food-corpus-b-FROZEN-20260712-324f4ab3-c63-20260802

Supposed to	N	Correct	Wrong	Unverified
▸ LOG — log the entry	86	59 (69%)	27 (31%)	0 (0%)
27 handled wrong — click one to jump to its full detail below "I ate 2 slices sourdough toast with 14 grams butter." WRITE-TRUTH FAIL — IMPLAUSIBLE: 530 kcal is implausibly HIGH for 2 count of Toast (typical 160-260 kcal [per-serving]) "Breakfast was 45 grams dry rolled oats cooked with water." OVER-ASK — asked instead of logging (no saved row). "Add one large navel orange." OVER-ASK — asked instead of logging (no saved row). "Dinner was 5 ounces cod plus 160 grams roasted potatoes." OVER-ASK — asked instead of logging (no saved row). "Track 7 ounces rotisserie chicken, meat only." WRITE-TRUTH FAIL — WRONG/MISSING item "7 ounces rotisserie chicken meat only" — no saved row with matching identity (rows: Chicken breast, Pork, Cured, Ham with Natural Juices, Spiral Slice, Meat Only, Boneless, Separable Lean Only, Heated, Roasted) "Log 250 milliliters whole milk." OVER-ASK — asked instead of logging (no saved row). "Record 100 grams cooked pasta with 125 grams marinara." OVER-ASK — asked instead of logging (no saved row). "Add 140 grams sweet potato mash." OVER-ASK — asked instead of logging (no saved row). "Add 1 cup canned corn drained." OVER-ASK — asked instead of logging (no saved row). "Add 1 cup beef chili." OVER-ASK — asked instead of logging (no saved row). "I had 1 cup cooked couscous." OVER-ASK — asked instead of logging (no saved row). "Track 6 ounces tilapia with lemon, no oil." Reported could-not-confirm (write-confirmation false negative). "Record 45 grams pita chips." WRITE-TRUTH FAIL — WRONG/MISSING item "45 grams pita chips" — no saved row with matching identity (rows: Potato Chips) "Record 2 cups cauliflower rice." OVER-ASK — asked instead of logging (no saved row). "Add 5 ounces lamb chop, cooked weight." OVER-ASK — asked instead of logging (no saved row). "Record one Fairlife Core Power 26 gram vanilla shake." WRITE-TRUTH FAIL — IMPLAUSIBLE: 15 kcal is implausibly LOW for 1 count of Fairlife Core Power Elite Vanilla Protein Shake (typical 160-250 kcal [per-serving]) "Add a Starbucks grande flat white with whole milk." OVER-ASK — asked instead of logging (no saved row). "Track a Chipotle chicken bowl: white rice, black beans, mild salsa, cheese, lettuce." OVER-ASK — asked instead of logging (no saved row). "Log a McDonalds Egg McMuffin." OVER-ASK — asked instead of logging (no saved row). "I ate a Taco Bell bean burrito." OVER-ASK — asked instead of logging (no saved row). "Track one Dave’s Killer Bread thin-sliced Good Seed slice." OVER-ASK — asked instead of logging (no saved row). "Log a Siggi’s strawberry skyr cup." OVER-ASK — asked instead of logging (no saved row). "I had a Panera Fuji apple salad with chicken, full size." OVER-ASK — asked instead of logging (no saved row). "Record one Celsius peach vibe can." OVER-ASK — asked instead of logging (no saved row). "Add a Kodiak Cakes protein waffle, one waffle." WRITE-TRUTH FAIL — WRONG item/product identity "Kodiak Cakes protein waffle" — saved row does not hit product alias (rows: Birthday Cake Power Waffles With Sprinkles, Waffle, Plain, Waffle, Plain) "Track a Subway six inch turkey on wheat with provolone." OVER-ASK — asked instead of logging (no saved row). "Log one KIND dark chocolate nuts and sea salt bar." WRITE-TRUTH FAIL — WRONG item/product identity "KIND dark chocolate nuts" — saved row does not hit product alias (rows: Dark chocolate, No Cow Chocolate Sea Salt Bar)
Total	86	59 (69%)	27	0

PASSshould log the entry"Log 170 grams plain nonfat Greek yogurt." (difficulty —)0.3s

Verdict Expected LOG — should log the entry. PASS: Logged (write-truth verified): Nonfat Greek yogurt, Nonfat Greek yogurt — card not captured.

Why verdict Logged (write-truth verified): Nonfat Greek yogurt, Nonfat Greek yogurt — card not captured.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.3s

1 · TTS said Logged 170 grams plain nonfat greek yogurt.

2 · Card shown Logged 170 grams plain nonfat greek yogurt.

3 · App data rows written created food_log_entry: Nonfat Greek yogurt ×1 (170 g) 100 cal · 17.5g P · 6.1g C · 0.7g F
created food_log_entry: Nonfat Greek yogurt ×1 (170 g) 100 cal · 17.5g P · 6.1g C · 0.7g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:05:06.204Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched the common-food list to "Nonfat Greek yogurt" at 90% confidence.
Matched your past food logs to "Plain Greek Yogurt" at 82% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 0.4 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

FAILshould log the entry"I ate 2 slices sourdough toast with 14 grams butter." (difficulty —)0.4s

Verdict Expected LOG — should log the entry. FAIL: WRITE-TRUTH FAIL — IMPLAUSIBLE: 530 kcal is implausibly HIGH for 2 count of Toast (typical 160-260 kcal [per-serving])

Why verdict WRITE-TRUTH FAIL — IMPLAUSIBLE: 530 kcal is implausibly HIGH for 2 count of Toast (typical 160-260 kcal [per-serving])

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.4s

1 · TTS said Logged Toast and 14 grams butter. Assumed catalog default servings where you did not say an amount — tell me if that is not right.

2 · Card shown Logged Toast and 14 grams butter. Assumed catalog default servings where you did not say an amount — tell me if that is not right.

3 · App data rows written created food_log_entry: Toast ×2 (100 g) 530 cal · 18g P · 98g C · 6.4g F
created food_log_entry: Butter ×1 (14 g) 100 cal · 0.1g P · 0g C · 11.4g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:05:17.760Z

6 · Why this food No food-decision explanation was captured for this path.

7 · Lookup method No lookup method was captured for this path.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould log the entry"Record 3.5 ounces roasted turkey breast." (difficulty —)0.2s

Verdict Expected LOG — should log the entry. PASS: Logged (write-truth verified): Turkey breast — card not captured.

Why verdict Logged (write-truth verified): Turkey breast — card not captured.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.2s

1 · TTS said Logged 3.5 ounces roasted turkey breast.

2 · Card shown Logged 3.5 ounces roasted turkey breast.

3 · App data rows written created food_log_entry: Turkey breast ×1 (3.5 oz) 134 cal · 28.8g P · 0g C · 1.7g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:05:29.057Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched the common-food list to "Turkey breast" at 90% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 0.5 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

FAILshould log the entry"Breakfast was 45 grams dry rolled oats cooked with water." (difficulty —)2.2s

Verdict Expected LOG — should log the entry. FAIL: OVER-ASK — asked instead of logging (no saved row).

Why verdict OVER-ASK — asked instead of logging (no saved row).

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 2.2s

1 · TTS said What exact food and amount should I use for Dry oats? I did not log it yet because I could not safely finish that food log.

2 · Card shown What exact food and amount should I use for Dry oats? I did not log it yet because I could not safely finish that food log.

3 · App data rows written No food/workout/list/note data row captured for this utterance.This does not mean "nothing anywhere was stored"; spoken/chat evidence is shown above, and some chat/session records may live outside this write-capture field. It means the report did not capture an app data-row write such as food_log_entries or workout_sets.

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:05:42.433Z

6 · Why this food No food-decision explanation was captured for this path.

7 · Lookup method No lookup method was captured for this path.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

FAILshould log the entry"Add one large navel orange." (difficulty —)5.4s

Verdict Expected LOG — should log the entry. FAIL: OVER-ASK — asked instead of logging (no saved row).

Why verdict OVER-ASK — asked instead of logging (no saved row).

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 5.4s

1 · TTS said What exact food and amount should I use for Orange? I did not log it yet because I could not safely finish that food log.

2 · Card shown What exact food and amount should I use for Orange? I did not log it yet because I could not safely finish that food log.

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:05:58.971Z

6 · Why this food No food-decision explanation was captured for this path.

7 · Lookup method No lookup method was captured for this path.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould log the entry"Track 220 grams cooked jasmine rice." (difficulty —)0.2s

Verdict Expected LOG — should log the entry. PASS: Logged (write-truth verified): Cooked jasmine rice — card not captured.

Why verdict Logged (write-truth verified): Cooked jasmine rice — card not captured.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.2s

1 · TTS said Logged 220 grams cooked jasmine rice.

2 · Card shown Logged 220 grams cooked jasmine rice.

3 · App data rows written created food_log_entry: Cooked jasmine rice ×1 (220 g) 286 cal · 5.3g P · 62g C · 0.7g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:06:10.291Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched the common-food list to "Cooked jasmine rice" at 90% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 0.5 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould log the entry"For lunch I had 6 ounces grilled flank steak." (difficulty —)0.4s

Verdict Expected LOG — should log the entry. PASS: Logged (write-truth verified): Grilled Steak — card not captured.

Why verdict Logged (write-truth verified): Grilled Steak — card not captured.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.4s

1 · TTS said Logged Grilled Steak. Using your recent Grilled Steak history. Tell me if that is wrong.

2 · Card shown Logged Grilled Steak. Using your recent Grilled Steak history. Tell me if that is wrong.

3 · App data rows written created food_log_entry: Grilled Steak ×1 (ounce) 340 cal · 42g P · 0g C · 18g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:06:21.813Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched your past food logs to "Grilled Steak" at 82% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 0.5 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould log the entry"Put down 85 grams avocado." (difficulty —)0.2s

Verdict Expected LOG — should log the entry. PASS: Logged (write-truth verified): Avocado — card not captured.

Why verdict Logged (write-truth verified): Avocado — card not captured.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.2s

1 · TTS said Logged 85 grams avocado.

2 · Card shown Logged 85 grams avocado.

3 · App data rows written created food_log_entry: Avocado ×1 (85 g) 136 cal · 1.7g P · 7.2g C · 12.5g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:06:33.149Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched the common-food list to "Avocado" at 90% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 0.4 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould log the entry"I had a cup and a half of cooked lentils." (difficulty —)0.4s

Verdict Expected LOG — should log the entry. PASS: Logged (write-truth verified): Lentils — card not captured.

Why verdict Logged (write-truth verified): Lentils — card not captured.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.4s

1 · TTS said Logged a cup and a half of cooked lentils. Assumed 1.5 cup as the default serving. Tell me if you want a different amount.

2 · Card shown Logged a cup and a half of cooked lentils. Assumed 1.5 cup as the default serving. Tell me if you want a different amount.

3 · App data rows written created food_log_entry: Lentils ×1 (1.5 cup) 345 cal · 26.7g P · 59.7g C · 1.2g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:06:44.716Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched the common-food list to "Lentils" at 90% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 0.6 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould log the entry"Record 28 grams cheddar cheese." (difficulty —)0.2s

Verdict Expected LOG — should log the entry. PASS: Logged (write-truth verified): Cheddar cheese — card not captured.

Why verdict Logged (write-truth verified): Cheddar cheese — card not captured.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.2s

1 · TTS said Logged 28 grams cheddar cheese.

2 · Card shown Logged 28 grams cheddar cheese.

3 · App data rows written created food_log_entry: Cheddar cheese ×1 (28 g) 113 cal · 6.4g P · 1g C · 9.3g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:06:56.108Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched the common-food list to "Cheddar cheese" at 90% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 0.5 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould log the entry"Add 12 baby carrots and 2 tablespoons hummus." (difficulty —)2.2s

Verdict Expected LOG — should log the entry. PASS: Logged (write-truth verified): Baby carrots, Hummus, Plain — card not captured.

Why verdict Logged (write-truth verified): Baby carrots, Hummus, Plain — card not captured.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 2.2s

1 · TTS said Logged 12 baby carrots and Hummus, Plain.

2 · Card shown Logged 12 baby carrots and Hummus, Plain.

3 · App data rows written created food_log_entry: Baby carrots ×1 (12 baby carrots) 42 cal · 0.7g P · 9.8g C · 0.1g F
created food_log_entry: Hummus, Plain ×1 (2 tablespoons (30 g)) 73 cal · 2.2g P · 4.5g C · 5.1g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:07:09.387Z

6 · Why this food No food-decision explanation was captured for this path.

7 · Lookup method No lookup method was captured for this path.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould log the entry"Track a 9 ounce baked potato, plain." (difficulty —)0.8s

Verdict Expected LOG — should log the entry. PASS: Logged (write-truth verified): Potato — card not captured.

Why verdict Logged (write-truth verified): Potato — card not captured.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.8s

1 · TTS said Logged a 9 ounce baked potato plain.

2 · Card shown Logged a 9 ounce baked potato plain.

3 · App data rows written created food_log_entry: Potato ×1 (9 oz) 222 cal · 4.8g P · 51.3g C · 0.3g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:07:21.293Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched the common-food list to "Potato" at 90% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 0.7 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

FAILshould log the entry"Dinner was 5 ounces cod plus 160 grams roasted potatoes." (difficulty —)6.0s

Verdict Expected LOG — should log the entry. FAIL: OVER-ASK — asked instead of logging (no saved row).

Why verdict OVER-ASK — asked instead of logging (no saved row).

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 6.0s

1 · TTS said What exact food and amount should I use for Cod? I did not log it yet because I could not safely finish that food log.

2 · Card shown What exact food and amount should I use for Cod? I did not log it yet because I could not safely finish that food log.

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:07:38.513Z

6 · Why this food No food-decision explanation was captured for this path.

7 · Lookup method No lookup method was captured for this path.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould log the entry"Log one cup cooked quinoa." (difficulty —)0.2s

Verdict Expected LOG — should log the entry. PASS: Logged (write-truth verified): Cooked quinoa — card not captured.

Why verdict Logged (write-truth verified): Cooked quinoa — card not captured.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.2s

1 · TTS said Logged one cup cooked quinoa. Assumed 1 cup as the default serving. Tell me if you want a different amount.

2 · Card shown Logged one cup cooked quinoa. Assumed 1 cup as the default serving. Tell me if you want a different amount.

3 · App data rows written created food_log_entry: Cooked quinoa ×1 (1 cup) 222 cal · 8.1g P · 39.4g C · 3.5g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:07:49.849Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched the common-food list to "Cooked quinoa" at 90% confidence.
Matched your past food logs to "Cooked Oatmeal" at 82% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 0.5 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould log the entry"Add 2 cups romaine and 30 grams parmesan." (difficulty —)0.3s

Verdict Expected LOG — should log the entry. PASS: Logged (write-truth verified): Parmesan cheese, Romaine lettuce — card not captured.

Why verdict Logged (write-truth verified): Parmesan cheese, Romaine lettuce — card not captured.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.3s

1 · TTS said Logged 2 cups romaine and 30 grams parmesan.

2 · Card shown Logged 2 cups romaine and 30 grams parmesan.

3 · App data rows written created food_log_entry: Romaine lettuce ×1 (2 cup) 16 cal · 1.1g P · 3.1g C · 0.3g F
created food_log_entry: Parmesan cheese ×1 (30 g) 124 cal · 9g P · 2.5g C · 8.6g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:08:01.249Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched the common-food list to "Parmesan cheese" at 90% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 0.7 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould log the entry"I ate 150 grams blueberries." (difficulty —)0.2s

Verdict Expected LOG — should log the entry. PASS: Logged (write-truth verified): Blueberries — card not captured.

Why verdict Logged (write-truth verified): Blueberries — card not captured.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.2s

1 · TTS said Logged 150 grams blueberries.

2 · Card shown Logged 150 grams blueberries.

3 · App data rows written created food_log_entry: Blueberries ×1 (150 g) 86 cal · 1g P · 21.8g C · 0.5g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:08:12.525Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched the common-food list to "Blueberries" at 90% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 0.5 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould log the entry"Record three scrambled eggs cooked with a teaspoon olive oil." (difficulty —)0.5s

Verdict Expected LOG — should log the entry. PASS: Logged (write-truth verified): Olive oil, Egg — card not captured.

Why verdict Logged (write-truth verified): Olive oil, Egg — card not captured.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.5s

1 · TTS said Logged three scrambled eggs cooked and a teaspoon olive oil. Assumed 1 tsp as the default serving. Tell me if you want a different amount. Tell me if that is not right.

2 · Card shown Logged three scrambled eggs cooked and a teaspoon olive oil. Assumed 1 tsp as the default serving. Tell me if you want a different amount. Tell me if that is not right.

3 · App data rows written created food_log_entry: Egg ×1 (3 eggs) 215 cal · 18.9g P · 1g C · 14.3g F
created food_log_entry: Olive oil ×1 (1 tsp) 40 cal · 0g P · 0g C · 4.5g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:08:24.220Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched the common-food list to "Egg" at 90% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 1.0 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

FAILshould log the entry"Track 7 ounces rotisserie chicken, meat only." (difficulty —)0.7s

Verdict Expected LOG — should log the entry. FAIL: WRITE-TRUTH FAIL — WRONG/MISSING item "7 ounces rotisserie chicken meat only" — no saved row with matching identity (rows: Chicken breast, Pork, Cured, Ham with Natural Juices, Spiral Slice, Meat Only, Boneless, Separable Lean Only, Heated, Roasted)

Why verdict WRITE-TRUTH FAIL — WRONG/MISSING item "7 ounces rotisserie chicken meat only" — no saved row with matching identity (rows: Chicken breast, Pork, Cured, Ham with Natural Juices, Spiral Slice, Meat Only, Boneless, Separable Lean Only, Heated, Roasted)

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.7s

1 · TTS said Logged Chicken breast and Pork, Cured, Ham with Natural Juices, Spiral Slice, Meat Only, Boneless, Separable Lean Only, Heated, Roasted. Assumed a typical serving of Pork, Cured, Ham with Natural Juices, Spiral Slice, Meat Only, Boneless, Separable Lean Only, Heated, Roasted while logging the rest of this meal. Tell me if that is not right.

2 · Card shown Logged Chicken breast and Pork, Cured, Ham with Natural Juices, Spiral Slice, Meat Only, Boneless, Separable Lean Only, Heated, Roasted. Assumed a typical serving of Pork, Cured, Ham with Natural Juices, Spiral Slice, Meat Only, Boneless, Separable Lean Only, Heated, Roasted while logging the rest of this meal. Tell me if that is not right.

3 · App data rows written created food_log_entry: Chicken breast ×1 (7 ounces (198.4 g)) 327 cal · 61.5g P · 0g C · 7.1g F
created food_log_entry: Pork, Cured, Ham with Natural Juices, Spiral Slice, Meat Only, Boneless, Separable Lean Only, Heated, Roasted ×1 (3 serving) 107 cal · 19.2g P · 0.9g C · 3.2g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:08:36.008Z

6 · Why this food No food-decision explanation was captured for this path.

7 · Lookup method No lookup method was captured for this path.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould log the entry"Snack was 32 grams almonds." (difficulty —)0.2s

Verdict Expected LOG — should log the entry. PASS: Logged (write-truth verified): Almonds — card not captured.

Why verdict Logged (write-truth verified): Almonds — card not captured.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.2s

1 · TTS said Logged 32 grams almonds.

2 · Card shown Logged 32 grams almonds.

3 · App data rows written created food_log_entry: Almonds ×1 (32 g) 185 cal · 6.8g P · 6.9g C · 16g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:08:47.322Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched the common-food list to "Almonds" at 90% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 0.6 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

FAILshould log the entry"Log 250 milliliters whole milk." (difficulty —)0.1s

Verdict Expected LOG — should log the entry. FAIL: OVER-ASK — asked instead of logging (no saved row).

Why verdict OVER-ASK — asked instead of logging (no saved row).

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.1s

1 · TTS said I couldn't convert 250 milliliters of Whole milk from the serving information I found. What does the label say for grams or calories? I did not log it yet.

2 · Card shown I couldn't convert 250 milliliters of Whole milk from the serving information I found. What does the label say for grams or calories? I did not log it yet.

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:08:58.594Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched your past food logs to "Coffee shop default Whole Milk Latte" at 82% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 0.7 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould log the entry"I had 1 tablespoon peanut butter on a medium banana." (difficulty —)0.4s

Verdict Expected LOG — should log the entry. PASS: Logged (write-truth verified): Peanut butter, Banana — card not captured.

Why verdict Logged (write-truth verified): Peanut butter, Banana — card not captured.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.4s

1 · TTS said Logged 1 tablespoon peanut butter and a medium banana. Assumed 1 medium banana as the default serving. Tell me if you want a different amount. Tell me if that is not right.

2 · Card shown Logged 1 tablespoon peanut butter and a medium banana. Assumed 1 medium banana as the default serving. Tell me if you want a different amount. Tell me if that is not right.

3 · App data rows written created food_log_entry: Peanut butter ×1 (1 tbsp) 88 cal · 3.8g P · 2.9g C · 7.6g F
created food_log_entry: Banana ×1 (1 medium banana) 105 cal · 1.3g P · 27.1g C · 0.4g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:09:21.641Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched the common-food list to "Peanut butter" at 90% confidence.
Matched your saved foods and aliases to "Medium Banana" at 60% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 1.1 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

FAILshould log the entry"Record 100 grams cooked pasta with 125 grams marinara." (difficulty —)1.0s

Verdict Expected LOG — should log the entry. FAIL: OVER-ASK — asked instead of logging (no saved row).

Why verdict OVER-ASK — asked instead of logging (no saved row).

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 1.0s

1 · TTS said I found Cooked pasta, but I also heard 125 marinara. Should I log those separately too?

2 · Card shown I found Cooked pasta, but I also heard 125 marinara. Should I log those separately too?

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:09:33.807Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched the common-food list to "Cooked pasta" at 90% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 1.1 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould log the entry"Add 4 ounces tofu and 2 cups broccoli." (difficulty —)0.3s

Verdict Expected LOG — should log the entry. PASS: Logged (write-truth verified): Firm tofu, Broccoli — card not captured.

Why verdict Logged (write-truth verified): Firm tofu, Broccoli — card not captured.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.3s

1 · TTS said Logged 4 ounces tofu and 2 cups broccoli.

2 · Card shown Logged 4 ounces tofu and 2 cups broccoli.

3 · App data rows written created food_log_entry: Firm tofu ×1 (4 oz) 163 cal · 19.3g P · 3.2g C · 9.9g F
created food_log_entry: Broccoli ×1 (2 cup) 64 cal · 4.4g P · 13.1g C · 0.7g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:09:45.227Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched the common-food list to "Broccoli" at 90% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 0.9 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould log the entry"Track 1 cup black beans drained." (difficulty —)0.2s

Verdict Expected LOG — should log the entry. PASS: Logged (write-truth verified): Black beans — card not captured.

Why verdict Logged (write-truth verified): Black beans — card not captured.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.2s

1 · TTS said Logged 1 cup black beans drained.

2 · Card shown Logged 1 cup black beans drained.

3 · App data rows written created food_log_entry: Black beans ×1 (1 cup) 227 cal · 15.3g P · 40.8g C · 0.9g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:09:56.524Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched the common-food list to "Black beans" at 90% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 1.0 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould log the entry"Put down 60 grams granola." (difficulty —)0.2s

Verdict Expected LOG — should log the entry. PASS: Logged (write-truth verified): Granola — card not captured.

Why verdict Logged (write-truth verified): Granola — card not captured.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.2s

1 · TTS said Logged 60 grams granola.

2 · Card shown Logged 60 grams granola.

3 · App data rows written created food_log_entry: Granola ×1 (60 g) 283 cal · 6g P · 38.4g C · 12g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:10:07.850Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched the common-food list to "Granola" at 90% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 0.7 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould log the entry"I ate 180 grams cottage cheese." (difficulty —)0.2s

Verdict Expected LOG — should log the entry. PASS: Logged (write-truth verified): Cottage cheese — card not captured.

Why verdict Logged (write-truth verified): Cottage cheese — card not captured.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.2s

1 · TTS said Logged 180 grams cottage cheese.

2 · Card shown Logged 180 grams cottage cheese.

3 · App data rows written created food_log_entry: Cottage cheese ×1 (180 g) 176 cal · 20g P · 6.1g C · 7.7g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:10:19.139Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched the common-food list to "Cottage cheese" at 90% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 0.9 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould log the entry"Log one medium pear and 18 grams walnuts." (difficulty —)0.3s

Verdict Expected LOG — should log the entry. PASS: Logged (write-truth verified): Walnuts, Pear — card not captured.

Why verdict Logged (write-truth verified): Walnuts, Pear — card not captured.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.3s

1 · TTS said Logged one medium pear and 18 grams walnuts. Assumed 1 pear as the default serving. Tell me if you want a different amount. Tell me if that is not right.

2 · Card shown Logged one medium pear and 18 grams walnuts. Assumed 1 pear as the default serving. Tell me if you want a different amount. Tell me if that is not right.

3 · App data rows written created food_log_entry: Pear ×1 (1 pear) 101 cal · 0.7g P · 27.1g C · 0.2g F
created food_log_entry: Walnuts ×1 (18 g) 118 cal · 2.7g P · 2.5g C · 11.7g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:10:30.534Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched the common-food list to "Walnuts" at 90% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 1.0 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould log the entry"Record 2 ounces smoked salmon." (difficulty —)0.2s

Verdict Expected LOG — should log the entry. PASS: Logged (write-truth verified): Smoked salmon — card not captured.

Why verdict Logged (write-truth verified): Smoked salmon — card not captured.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.2s

1 · TTS said Logged 2 ounces smoked salmon.

2 · Card shown Logged 2 ounces smoked salmon.

3 · App data rows written created food_log_entry: Smoked salmon ×1 (2 oz) 66 cal · 10.4g P · 0g C · 2.4g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:10:41.838Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched the common-food list to "Smoked salmon" at 90% confidence.
Matched your past food logs to "Baked Salmon" at 82% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 0.8 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould log the entry"Add 1 cup cooked farro." (difficulty —)0.2s

Verdict Expected LOG — should log the entry. PASS: Logged (write-truth verified): Cooked farro — card not captured.

Why verdict Logged (write-truth verified): Cooked farro — card not captured.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.2s

1 · TTS said Logged 1 cup cooked farro.

2 · Card shown Logged 1 cup cooked farro.

3 · App data rows written created food_log_entry: Cooked farro ×1 (1 cup) 240 cal · 8.6g P · 50.7g C · 1.6g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:10:53.156Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched the common-food list to "Cooked farro" at 90% confidence.
Matched your past food logs to "Cooked quinoa" at 82% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 0.9 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould log the entry"Track 90 grams cucumber and 90 grams tomato." (difficulty —)0.3s

Verdict Expected LOG — should log the entry. PASS: Logged (write-truth verified): Cucumber, Tomato — card not captured.

Why verdict Logged (write-truth verified): Cucumber, Tomato — card not captured.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.3s

1 · TTS said Logged 90 grams cucumber and 90 grams tomato.

2 · Card shown Logged 90 grams cucumber and 90 grams tomato.

3 · App data rows written created food_log_entry: Cucumber ×1 (90 g) 14 cal · 0.6g P · 3.2g C · 0.1g F
created food_log_entry: Tomato ×1 (90 g) 16 cal · 0.8g P · 3.5g C · 0.2g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:11:04.574Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched the common-food list to "Cucumber" at 90% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 1.2 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould log the entry"Log 6 ounces pork tenderloin." (difficulty —)0.2s

Verdict Expected LOG — should log the entry. PASS: Logged (write-truth verified): Pork tenderloin — card not captured.

Why verdict Logged (write-truth verified): Pork tenderloin — card not captured.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.2s

1 · TTS said Logged 6 ounces pork tenderloin.

2 · Card shown Logged 6 ounces pork tenderloin.

3 · App data rows written created food_log_entry: Pork tenderloin ×1 (6 oz) 243 cal · 44.2g P · 0g C · 6g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:11:15.920Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched the common-food list to "Pork tenderloin" at 90% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 1.0 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould log the entry"Record 1 cup edamame shelled." (difficulty —)0.2s

Verdict Expected LOG — should log the entry. PASS: Logged (write-truth verified): Edamame — card not captured.

Why verdict Logged (write-truth verified): Edamame — card not captured.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.2s

1 · TTS said Logged 1 cup edamame shelled.

2 · Card shown Logged 1 cup edamame shelled.

3 · App data rows written created food_log_entry: Edamame ×1 (1 cup) 188 cal · 18.4g P · 13.8g C · 8.1g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:11:27.316Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched the common-food list to "Edamame" at 90% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 1.0 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould log the entry"Add a small flour tortilla with 50 grams grilled chicken." (difficulty —)0.5s

Verdict Expected LOG — should log the entry. PASS: Logged (write-truth verified): Flour tortilla, Chicken breast — card not captured.

Why verdict Logged (write-truth verified): Flour tortilla, Chicken breast — card not captured.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.5s

1 · TTS said Logged a small flour tortilla and 50 grams grilled chicken. Assumed 1 small flour tortilla as the default serving. Tell me if you want a different amount. Tell me if that is not right.

2 · Card shown Logged a small flour tortilla and 50 grams grilled chicken. Assumed 1 small flour tortilla as the default serving. Tell me if you want a different amount. Tell me if that is not right.

3 · App data rows written created food_log_entry: Flour tortilla ×1 (1 small flour tortilla) 74 cal · 2g P · 12g C · 1.9g F
created food_log_entry: Chicken breast ×1 (50 g) 83 cal · 15.5g P · 0g C · 1.8g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:11:39.018Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched the common-food list to "Flour tortilla" at 90% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 1.5 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould log the entry"I had 40 grams raisins." (difficulty —)0.2s

Verdict Expected LOG — should log the entry. PASS: Logged (write-truth verified): Raisins — card not captured.

Why verdict Logged (write-truth verified): Raisins — card not captured.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.2s

1 · TTS said Logged 40 grams raisins.

2 · Card shown Logged 40 grams raisins.

3 · App data rows written created food_log_entry: Raisins ×1 (40 g) 120 cal · 1.2g P · 31.7g C · 0.2g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:11:50.342Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched the common-food list to "Raisins" at 90% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 0.9 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould log the entry"Track 2 tablespoons chia seeds." (difficulty —)0.2s

Verdict Expected LOG — should log the entry. PASS: Logged (write-truth verified): Chia seeds — card not captured.

Why verdict Logged (write-truth verified): Chia seeds — card not captured.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.2s

1 · TTS said Logged 2 tablespoons chia seeds.

2 · Card shown Logged 2 tablespoons chia seeds.

3 · App data rows written created food_log_entry: Chia seeds ×1 (2 tbsp) 102 cal · 3.5g P · 8.8g C · 6.4g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:12:01.657Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched the common-food list to "Chia seeds" at 90% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 1.1 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould log the entry"Log 5 ounces shrimp cooked." (difficulty —)0.2s

Verdict Expected LOG — should log the entry. PASS: Logged (write-truth verified): Shrimp — card not captured.

Why verdict Logged (write-truth verified): Shrimp — card not captured.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.2s

1 · TTS said Logged 5 ounces shrimp cooked.

2 · Card shown Logged 5 ounces shrimp cooked.

3 · App data rows written created food_log_entry: Shrimp ×1 (5 oz) 140 cal · 34g P · 0.3g C · 0.4g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:12:12.965Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched the common-food list to "Shrimp" at 90% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 1.1 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould log the entry"Record 1 cup unsweetened applesauce." (difficulty —)0.2s

Verdict Expected LOG — should log the entry. PASS: Logged (write-truth verified): Unsweetened applesauce — card not captured.

Why verdict Logged (write-truth verified): Unsweetened applesauce — card not captured.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.2s

1 · TTS said Logged 1 cup unsweetened applesauce.

2 · Card shown Logged 1 cup unsweetened applesauce.

3 · App data rows written created food_log_entry: Unsweetened applesauce ×1 (1 cup) 102 cal · 0.5g P · 27.6g C · 0.2g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:12:24.293Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched the common-food list to "Unsweetened applesauce" at 90% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 1.2 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

FAILshould log the entry"Add 140 grams sweet potato mash." (difficulty —)4.0s

Verdict Expected LOG — should log the entry. FAIL: OVER-ASK — asked instead of logging (no saved row).

Why verdict OVER-ASK — asked instead of logging (no saved row).

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 4.0s

1 · TTS said Which one did you mean? 1. Sweet Potato Mash — T.G.I. FRIDAY'S, 2. Sweet Potato Mash — HUNGRY JACK, 3. Sweet Potato Mash — Basic American Foods, Inc.

2 · Card shown Which one did you mean? 1. Sweet Potato Mash — T.G.I. FRIDAY'S, 2. Sweet Potato Mash — HUNGRY JACK, 3. Sweet Potato Mash — Basic American Foods, Inc.

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:12:39.510Z

6 · Why this food No food-decision explanation was captured for this path.

7 · Lookup method No lookup method was captured for this path.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould log the entry"I ate 25 grams dark chocolate." (difficulty —)0.2s

Verdict Expected LOG — should log the entry. PASS: Logged (write-truth verified): Dark chocolate — card not captured.

Why verdict Logged (write-truth verified): Dark chocolate — card not captured.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.2s

1 · TTS said Logged 25 grams dark chocolate.

2 · Card shown Logged 25 grams dark chocolate.

3 · App data rows written created food_log_entry: Dark chocolate ×1 (25 g) 138 cal · 1.3g P · 15g C · 8.1g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:12:50.834Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched the common-food list to "Dark chocolate" at 90% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 1.2 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould log the entry"Track 2 cups air-popped popcorn." (difficulty —)0.3s

Verdict Expected LOG — should log the entry. PASS: Logged (write-truth verified): Air-popped popcorn — card not captured.

Why verdict Logged (write-truth verified): Air-popped popcorn — card not captured.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.3s

1 · TTS said Logged 2 cups air popped popcorn.

2 · Card shown Logged 2 cups air popped popcorn.

3 · App data rows written created food_log_entry: Air-popped popcorn ×1 (2 cup) 62 cal · 2.1g P · 12.5g C · 0.7g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:13:02.374Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched the common-food list to "Air-popped popcorn" at 90% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 1.5 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould log the entry"Log 1.25 cups low fat kefir." (difficulty —)0.2s

Verdict Expected LOG — should log the entry. PASS: Logged (write-truth verified): Kefir — card not captured.

Why verdict Logged (write-truth verified): Kefir — card not captured.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.2s

1 · TTS said Logged 1.25 cups low fat kefir.

2 · Card shown Logged 1.25 cups low fat kefir.

3 · App data rows written created food_log_entry: Kefir ×1 (1.25 cup) 126 cal · 10.1g P · 13.8g C · 3.1g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:13:25.813Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched the common-food list to "Kefir" at 90% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 1.4 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould log the entry"Record 4 ounces tempeh." (difficulty —)0.2s

Verdict Expected LOG — should log the entry. PASS: Logged (write-truth verified): Tempeh — card not captured.

Why verdict Logged (write-truth verified): Tempeh — card not captured.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.2s

1 · TTS said Logged 4 ounces tempeh.

2 · Card shown Logged 4 ounces tempeh.

3 · App data rows written created food_log_entry: Tempeh ×1 (4 oz) 218 cal · 23g P · 8.6g C · 12.2g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:13:37.170Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched the common-food list to "Tempeh" at 90% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 1.0 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

FAILshould log the entry"Add 1 cup canned corn drained." (difficulty —)5.4s

Verdict Expected LOG — should log the entry. FAIL: OVER-ASK — asked instead of logging (no saved row).

Why verdict OVER-ASK — asked instead of logging (no saved row).

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 5.4s

1 · TTS said What exact food and amount should I use for 1 cup canned corn drained? I did not log it yet because I could not safely finish that food log.

2 · Card shown What exact food and amount should I use for 1 cup canned corn drained? I did not log it yet because I could not safely finish that food log.

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:13:53.740Z

6 · Why this food No food-decision explanation was captured for this path.

7 · Lookup method No lookup method was captured for this path.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould log the entry"I had 70 grams turkey pepperoni." (difficulty —)2.1s

Verdict Expected LOG — should log the entry. PASS: Logged (write-truth verified): APPLEGATE NATURALS* Uncured Turkey Pepperoni — card not captured.

Why verdict Logged (write-truth verified): APPLEGATE NATURALS* Uncured Turkey Pepperoni — card not captured.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 2.1s

1 · TTS said Got it, I logged APPLEGATE NATURALS Uncured Turkey Pepperoni (70 grams (70 g)).

2 · Card shown I logged APPLEGATE NATURALS Uncured Turkey Pepperoni (70 grams (70 g)) — let me know if that's not right.

3 · App data rows written created food_log_entry: APPLEGATE NATURALS* Uncured Turkey Pepperoni ×1 (70 grams (70 g)) 187 cal · 23.3g P · 0g C · 9.3g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:14:06.967Z

6 · Why this food No food-decision explanation was captured for this path.

7 · Lookup method No lookup method was captured for this path.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould log the entry"Track 3 cups spinach cooked down." (difficulty —)0.2s

Verdict Expected LOG — should log the entry. PASS: Logged (write-truth verified): Spinach — card not captured.

Why verdict Logged (write-truth verified): Spinach — card not captured.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.2s

1 · TTS said Logged 3 cups spinach cooked down.

2 · Card shown Logged 3 cups spinach cooked down.

3 · App data rows written created food_log_entry: Spinach ×1 (3 cup) 21 cal · 2.6g P · 3.2g C · 0.4g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:14:18.497Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched the common-food list to "Spinach" at 90% confidence.
Matched your past food logs to "Cooked farro" at 82% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 1.3 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould log the entry"Log 2 rice cakes with 1 tablespoon almond butter." (difficulty —)0.6s

Verdict Expected LOG — should log the entry. PASS: Logged (write-truth verified): Almond butter, Plain rice cakes — card not captured.

Why verdict Logged (write-truth verified): Almond butter, Plain rice cakes — card not captured.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.6s

1 · TTS said Logged 2 rice cakes and 1 tablespoon almond butter.

2 · Card shown Logged 2 rice cakes and 1 tablespoon almond butter.

3 · App data rows written created food_log_entry: Plain rice cakes ×1 (2 rice cakes) 70 cal · 1.4g P · 14.7g C · 0.5g F
created food_log_entry: Almond butter ×1 (1 tbsp) 98 cal · 3.4g P · 3g C · 9g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:14:30.084Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched the common-food list to "Plain rice cakes" at 90% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 1.7 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould log the entry"Record 120 grams grilled zucchini." (difficulty —)0.2s

Verdict Expected LOG — should log the entry. PASS: Logged (write-truth verified): Zucchini — card not captured.

Why verdict Logged (write-truth verified): Zucchini — card not captured.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.2s

1 · TTS said Logged 120 grams grilled zucchini.

2 · Card shown Logged 120 grams grilled zucchini.

3 · App data rows written created food_log_entry: Zucchini ×1 (120 g) 20 cal · 1.4g P · 3.7g C · 0.4g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:14:41.493Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched the common-food list to "Zucchini" at 90% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 1.4 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

FAILshould log the entry"Add 1 cup beef chili." (difficulty —)0.2s

Verdict Expected LOG — should log the entry. FAIL: OVER-ASK — asked instead of logging (no saved row).

Why verdict OVER-ASK — asked instead of logging (no saved row).

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.2s

1 · TTS said What food did you have?

2 · Card shown What food did you have?

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:14:52.846Z

6 · Why this food No food-decision explanation was captured for this path.

7 · Lookup method No lookup method was captured for this path.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould log the entry"I ate 4 ounces turkey meatballs." (difficulty —)0.2s

Verdict Expected LOG — should log the entry. PASS: Logged (write-truth verified): Turkey meatballs — card not captured.

Why verdict Logged (write-truth verified): Turkey meatballs — card not captured.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.2s

1 · TTS said Logged 4 ounces turkey meatballs.

2 · Card shown Logged 4 ounces turkey meatballs.

3 · App data rows written created food_log_entry: Turkey meatballs ×1 (4 oz) 210 cal · 20.4g P · 6.8g C · 11.3g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:15:04.165Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched the common-food list to "Turkey meatballs" at 90% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 1.4 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould log the entry"Track 125 grams mango." (difficulty —)0.2s

Verdict Expected LOG — should log the entry. PASS: Logged (write-truth verified): Mango — card not captured.

Why verdict Logged (write-truth verified): Mango — card not captured.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.2s

1 · TTS said Logged 125 grams mango.

2 · Card shown Logged 125 grams mango.

3 · App data rows written created food_log_entry: Mango ×1 (125 g) 75 cal · 1g P · 18.8g C · 0.5g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:15:15.497Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched the common-food list to "Mango" at 90% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 1.2 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould log the entry"Log one cup split pea soup." (difficulty —)2.2s

Verdict Expected LOG — should log the entry. PASS: Logged (write-truth verified): Split Pea Soup, Canned, Reduced Sodium, Prepared with Water or Ready-To Serve — card not captured.

Why verdict Logged (write-truth verified): Split Pea Soup, Canned, Reduced Sodium, Prepared with Water or Ready-To Serve — card not captured.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 2.2s

1 · TTS said Logged Split Pea Soup, Canned, Reduced Sodium, Prepared with Water or Ready-To Serve (one cup (245 g)) with an estimated macro range.

2 · Card shown Logged Split Pea Soup, Canned, Reduced Sodium, Prepared with Water or Ready-To Serve (one cup (245 g)) — the macros are an estimate since this food varies by brand.

3 · App data rows written created food_log_entry: Split Pea Soup, Canned, Reduced Sodium, Prepared with Water or Ready-To Serve ×1 (one cup (245 g)) 174 cal · 9.6g P · 28.9g C · 2.2g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:15:28.889Z

6 · Why this food No food-decision explanation was captured for this path.

7 · Lookup method No lookup method was captured for this path.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould log the entry"Record 3 ounces goat cheese." (difficulty —)0.2s

Verdict Expected LOG — should log the entry. PASS: Logged (write-truth verified): Goat cheese — card not captured.

Why verdict Logged (write-truth verified): Goat cheese — card not captured.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.2s

1 · TTS said Logged 3 ounces goat cheese.

2 · Card shown Logged 3 ounces goat cheese.

3 · App data rows written created food_log_entry: Goat cheese ×1 (3 oz) 310 cal · 18.4g P · 0.1g C · 25.3g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:15:40.206Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched the common-food list to "Goat cheese" at 90% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 1.4 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould log the entry"Add 2 cups watermelon cubes." (difficulty —)0.2s

Verdict Expected LOG — should log the entry. PASS: Logged (write-truth verified): Watermelon — card not captured.

Why verdict Logged (write-truth verified): Watermelon — card not captured.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.2s

1 · TTS said Logged 2 cups watermelon cubes.

2 · Card shown Logged 2 cups watermelon cubes.

3 · App data rows written created food_log_entry: Watermelon ×1 (2 cup) 91 cal · 1.8g P · 23.1g C · 0.6g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:15:51.549Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched the common-food list to "Watermelon" at 90% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 1.5 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

FAILshould log the entry"I had 1 cup cooked couscous." (difficulty —)2.7s

Verdict Expected LOG — should log the entry. FAIL: OVER-ASK — asked instead of logging (no saved row).

Why verdict OVER-ASK — asked instead of logging (no saved row).

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 2.7s

1 · TTS said What exact food and amount should I use for 1 cup cooked couscous? I did not log it yet because I could not safely finish that food log.

2 · Card shown What exact food and amount should I use for 1 cup cooked couscous? I did not log it yet because I could not safely finish that food log.

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:16:05.461Z

6 · Why this food No food-decision explanation was captured for this path.

7 · Lookup method No lookup method was captured for this path.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

FAILshould log the entry"Track 6 ounces tilapia with lemon, no oil." (difficulty —)0.1s

Verdict Expected LOG — should log the entry. FAIL: Reported could-not-confirm (write-confirmation false negative).

Why verdict Reported could-not-confirm (write-confirmation false negative).

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.1s

1 · TTS said I wasn't sure which food you meant to correct. Could you try again with a bit more detail?

2 · Card shown I wasn't sure which food you meant to correct. Could you try again with a bit more detail?

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:16:16.683Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched the common-food list to "Tilapia" at 90% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 1.6 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould log the entry"Log 75 grams sour cream." (difficulty —)0.2s

Verdict Expected LOG — should log the entry. PASS: Logged (write-truth verified): Sour cream — card not captured.

Why verdict Logged (write-truth verified): Sour cream — card not captured.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.2s

1 · TTS said Logged 75 grams sour cream.

2 · Card shown Logged 75 grams sour cream.

3 · App data rows written created food_log_entry: Sour cream ×1 (75 g) 147 cal · 2.3g P · 4.2g C · 13.5g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:16:27.976Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched the common-food list to "Sour cream" at 90% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 1.4 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould log the entry"Record 1 cup roasted Brussels sprouts." (difficulty —)0.2s

Verdict Expected LOG — should log the entry. PASS: Logged (write-truth verified): Brussels sprouts — card not captured.

Why verdict Logged (write-truth verified): Brussels sprouts — card not captured.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.2s

1 · TTS said Logged 1 cup roasted brussels sprouts.

2 · Card shown Logged 1 cup roasted brussels sprouts.

3 · App data rows written created food_log_entry: Brussels sprouts ×1 (1 cup) 70 cal · 5.4g P · 14.4g C · 0.5g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:16:39.318Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched the common-food list to "Brussels sprouts" at 90% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 1.7 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould log the entry"Add half cup ricotta." (difficulty —)0.2s

Verdict Expected LOG — should log the entry. PASS: Logged (write-truth verified): Ricotta cheese — card not captured.

Why verdict Logged (write-truth verified): Ricotta cheese — card not captured.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.2s

1 · TTS said Logged half cup ricotta. Assumed 0.5 cup as the default serving. Tell me if you want a different amount.

2 · Card shown Logged half cup ricotta. Assumed 0.5 cup as the default serving. Tell me if you want a different amount.

3 · App data rows written created food_log_entry: Ricotta cheese ×1 (0.5 cup) 182 cal · 11.8g P · 7.4g C · 11.7g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:16:50.647Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched the common-food list to "Ricotta cheese" at 90% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 1.4 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould log the entry"I ate 95 grams grapes." (difficulty —)0.2s

Verdict Expected LOG — should log the entry. PASS: Logged (write-truth verified): Grapes — card not captured.

Why verdict Logged (write-truth verified): Grapes — card not captured.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.2s

1 · TTS said Logged 95 grams grapes.

2 · Card shown Logged 95 grams grapes.

3 · App data rows written created food_log_entry: Grapes ×1 (95 g) 66 cal · 0.7g P · 17.2g C · 0.2g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:17:01.936Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched the common-food list to "Grapes" at 90% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 1.2 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould log the entry"Track 4 ounces seitan." (difficulty —)0.2s

Verdict Expected LOG — should log the entry. PASS: Logged (write-truth verified): Seitan — card not captured.

Why verdict Logged (write-truth verified): Seitan — card not captured.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.2s

1 · TTS said Logged 4 ounces seitan.

2 · Card shown Logged 4 ounces seitan.

3 · App data rows written created food_log_entry: Seitan ×1 (4 oz) 160 cal · 28g P · 6.7g C · 2.7g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:17:13.244Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched the common-food list to "Seitan" at 90% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 1.3 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould log the entry"Log 2 cups minestrone soup." (difficulty —)0.6s

Verdict Expected LOG — should log the entry. PASS: Logged (write-truth verified): Minestrone soup — card not captured.

Why verdict Logged (write-truth verified): Minestrone soup — card not captured.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.6s

1 · TTS said Logged 2 cups minestrone soup.

2 · Card shown Logged 2 cups minestrone soup.

3 · App data rows written created food_log_entry: Minestrone soup ×1 (2 cup) 191 cal · 8.8g P · 27.4g C · 5.4g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:17:36.717Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched the common-food list to "Minestrone soup" at 90% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 2.1 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

FAILshould log the entry"Record 45 grams pita chips." (difficulty —)0.3s

Verdict Expected LOG — should log the entry. FAIL: WRITE-TRUTH FAIL — WRONG/MISSING item "45 grams pita chips" — no saved row with matching identity (rows: Potato Chips)

Why verdict WRITE-TRUTH FAIL — WRONG/MISSING item "45 grams pita chips" — no saved row with matching identity (rows: Potato Chips)

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.3s

1 · TTS said Logged Potato Chips (45 grams (45 g)) with an estimated macro range.

2 · Card shown Logged Potato Chips (45 grams (45 g)) — the macros are an estimate since this food varies by brand.

3 · App data rows written created food_log_entry: Potato Chips ×1 (45 grams (45 g)) 241 cal · 3.2g P · 23.9g C · 15.8g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:17:48.110Z

6 · Why this food No food-decision explanation was captured for this path.

7 · Lookup method No lookup method was captured for this path.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould log the entry"Add 1 large kiwi." (difficulty —)0.2s

Verdict Expected LOG — should log the entry. PASS: Logged (write-truth verified): Kiwi — card not captured.

Why verdict Logged (write-truth verified): Kiwi — card not captured.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.2s

1 · TTS said Logged 1 large kiwi.

2 · Card shown Logged 1 large kiwi.

3 · App data rows written created food_log_entry: Kiwi ×1 (1 kiwi) 42 cal · 0.8g P · 10.1g C · 0.3g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:17:59.534Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched the common-food list to "Kiwi" at 90% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 1.4 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould log the entry"I had 3 ounces ham." (difficulty —)0.2s

Verdict Expected LOG — should log the entry. PASS: Logged (write-truth verified): Ham — card not captured.

Why verdict Logged (write-truth verified): Ham — card not captured.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.2s

1 · TTS said Logged 3 ounces ham.

2 · Card shown Logged 3 ounces ham.

3 · App data rows written created food_log_entry: Ham ×1 (3 oz) 123 cal · 17.9g P · 1.3g C · 4.7g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:18:12.303Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched the common-food list to "Ham" at 90% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 0.3 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould log the entry"Track 1 cup egg whites." (difficulty —)0.2s

Verdict Expected LOG — should log the entry. PASS: Logged (write-truth verified): Egg whites — card not captured.

Why verdict Logged (write-truth verified): Egg whites — card not captured.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.2s

1 · TTS said Logged 1 cup egg whites.

2 · Card shown Logged 1 cup egg whites.

3 · App data rows written created food_log_entry: Egg whites ×1 (1 cup) 126 cal · 26.5g P · 1.7g C · 0.5g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:18:23.634Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched the common-food list to "Egg whites" at 90% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 1.5 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould log the entry"Log 35 grams sunflower seeds." (difficulty —)0.2s

Verdict Expected LOG — should log the entry. PASS: Logged (write-truth verified): Sunflower seeds — card not captured.

Why verdict Logged (write-truth verified): Sunflower seeds — card not captured.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.2s

1 · TTS said Logged 35 grams sunflower seeds.

2 · Card shown Logged 35 grams sunflower seeds.

3 · App data rows written created food_log_entry: Sunflower seeds ×1 (35 g) 204 cal · 7.3g P · 7g C · 18g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:18:35.126Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched the common-food list to "Sunflower seeds" at 90% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 1.6 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

FAILshould log the entry"Record 2 cups cauliflower rice." (difficulty —)3.9s

Verdict Expected LOG — should log the entry. FAIL: OVER-ASK — asked instead of logging (no saved row).

Why verdict OVER-ASK — asked instead of logging (no saved row).

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 3.9s

1 · TTS said Which type of cauliflower rice was it? There are several with different nutrition.

2 · Card shown Which type of cauliflower rice was it? There are several with different nutrition.

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:18:50.029Z

6 · Why this food No food-decision explanation was captured for this path.

7 · Lookup method No lookup method was captured for this path.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

FAILshould log the entry"Add 5 ounces lamb chop, cooked weight." (difficulty —)3.2s

Verdict Expected LOG — should log the entry. FAIL: OVER-ASK — asked instead of logging (no saved row).

Why verdict OVER-ASK — asked instead of logging (no saved row).

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 3.2s

1 · TTS said I need to resolve cooked weight before I log this meal. What should I use for cooked weight?

2 · Card shown I need to resolve cooked weight before I log this meal. What should I use for cooked weight?

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:19:04.418Z

6 · Why this food No food-decision explanation was captured for this path.

7 · Lookup method No lookup method was captured for this path.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould log the entry"I ate 110 grams pineapple." (difficulty —)0.2s

Verdict Expected LOG — should log the entry. PASS: Logged (write-truth verified): Pineapple — card not captured.

Why verdict Logged (write-truth verified): Pineapple — card not captured.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.2s

1 · TTS said Logged 110 grams pineapple.

2 · Card shown Logged 110 grams pineapple.

3 · App data rows written created food_log_entry: Pineapple ×1 (110 g) 55 cal · 0.6g P · 14.4g C · 0.1g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:19:15.728Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched the common-food list to "Pineapple" at 90% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 1.5 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould log the entry"Track 1 cup baked beans." (difficulty —)0.2s

Verdict Expected LOG — should log the entry. PASS: Logged (write-truth verified): Baked beans — card not captured.

Why verdict Logged (write-truth verified): Baked beans — card not captured.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.2s

1 · TTS said Logged 1 cup baked beans.

2 · Card shown Logged 1 cup baked beans.

3 · App data rows written created food_log_entry: Baked beans ×1 (1 cup) 273 cal · 11.7g P · 56.2g C · 2.3g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:19:27.069Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched the common-food list to "Baked beans" at 90% confidence.
Matched your past food logs to "Black beans" at 82% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 1.6 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould log the entry"Log a Chobani Zero Sugar vanilla cup." (difficulty —)0.5s

Verdict Expected LOG — should log the entry. PASS: Logged (write-truth verified): Zero Sugar Vanilla Yogurt — card not captured.

Why verdict Logged (write-truth verified): Zero Sugar Vanilla Yogurt — card not captured.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.5s

1 · TTS said Got it, I logged Zero Sugar Vanilla Yogurt — Chobani.

2 · Card shown I logged Zero Sugar Vanilla Yogurt — Chobani — let me know if that's not right.

3 · App data rows written created food_log_entry: Zero Sugar Vanilla Yogurt ×1 (cup) 120 cal · 15g P · 9g C · 0g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:19:38.800Z

6 · Why this food No food-decision explanation was captured for this path.

7 · Lookup method No lookup method was captured for this path.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould log the entry"I had a Quest chocolate chip cookie dough protein bar." (difficulty —)0.6s

Verdict Expected LOG — should log the entry. PASS: Logged (write-truth verified): Quest Chocolate Chip Cookie Dough Protein Bar — card not captured.

Why verdict Logged (write-truth verified): Quest Chocolate Chip Cookie Dough Protein Bar — card not captured.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.6s

1 · TTS said Logged Quest Chocolate Chip Cookie Dough Protein Bar with an estimated macro range.

2 · Card shown Logged Quest Chocolate Chip Cookie Dough Protein Bar — the macros are an estimate since this food varies by brand.

3 · App data rows written created food_log_entry: Quest Chocolate Chip Cookie Dough Protein Bar ×1 (one) 190 cal · 21g P · 22g C · 9g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:19:50.607Z

6 · Why this food No food-decision explanation was captured for this path.

7 · Lookup method No lookup method was captured for this path.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

FAILshould log the entry"Record one Fairlife Core Power 26 gram vanilla shake." (difficulty —)1.6s

Verdict Expected LOG — should log the entry. FAIL: WRITE-TRUTH FAIL — IMPLAUSIBLE: 15 kcal is implausibly LOW for 1 count of Fairlife Core Power Elite Vanilla Protein Shake (typical 160-250 kcal [per-serving])

Why verdict WRITE-TRUTH FAIL — IMPLAUSIBLE: 15 kcal is implausibly LOW for 1 count of Fairlife Core Power Elite Vanilla Protein Shake (typical 160-250 kcal [per-serving])

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 1.6s

1 · TTS said Logged Fairlife Core Power Elite Vanilla Protein Shake (26 gram (26 g)) with an estimated macro range.

2 · Card shown Logged Fairlife Core Power Elite Vanilla Protein Shake (26 gram (26 g)) — the macros are an estimate since this food varies by brand.

3 · App data rows written created food_log_entry: Fairlife Core Power Elite Vanilla Protein Shake ×1 (26 gram (26 g)) 15 cal · 2.6g P · 0.5g C · 0.2g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:20:03.374Z

6 · Why this food No food-decision explanation was captured for this path.

7 · Lookup method No lookup method was captured for this path.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

FAILshould log the entry"Add a Starbucks grande flat white with whole milk." (difficulty —)5.4s

Verdict Expected LOG — should log the entry. FAIL: OVER-ASK — asked instead of logging (no saved row).

Why verdict OVER-ASK — asked instead of logging (no saved row).

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 5.4s

1 · TTS said What exact food and amount should I use for Whole milk? I did not log it yet because I could not safely finish that food log.

2 · Card shown What exact food and amount should I use for Whole milk? I did not log it yet because I could not safely finish that food log.

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:20:19.961Z

6 · Why this food No food-decision explanation was captured for this path.

7 · Lookup method No lookup method was captured for this path.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

FAILshould log the entry"Track a Chipotle chicken bowl: white rice, black beans, mild salsa, cheese, lettuce." (difficulty —)3.0s

Verdict Expected LOG — should log the entry. FAIL: OVER-ASK — asked instead of logging (no saved row).

Why verdict OVER-ASK — asked instead of logging (no saved row).

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 3.0s

1 · TTS said What should I use for lettuce?

2 · Card shown What should I use for lettuce?

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:20:28.868Z

6 · Why this food No food-decision explanation was captured for this path.

7 · Lookup method No lookup method was captured for this path.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

FAILshould log the entry"Log a McDonalds Egg McMuffin." (difficulty —)5.4s

Verdict Expected LOG — should log the entry. FAIL: OVER-ASK — asked instead of logging (no saved row).

Why verdict OVER-ASK — asked instead of logging (no saved row).

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 5.4s

1 · TTS said What exact food and amount should I use for Egg? I did not log it yet because I could not safely finish that food log.

2 · Card shown What exact food and amount should I use for Egg? I did not log it yet because I could not safely finish that food log.

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:20:50.696Z

6 · Why this food No food-decision explanation was captured for this path.

7 · Lookup method No lookup method was captured for this path.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

FAILshould log the entry"I ate a Taco Bell bean burrito." (difficulty —)2.7s

Verdict Expected LOG — should log the entry. FAIL: OVER-ASK — asked instead of logging (no saved row).

Why verdict OVER-ASK — asked instead of logging (no saved row).

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 2.7s

1 · TTS said What exact food and amount should I use for a taco bell bean burrito? I did not log it yet because I could not safely finish that food log.

2 · Card shown What exact food and amount should I use for a taco bell bean burrito? I did not log it yet because I could not safely finish that food log.

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:21:04.624Z

6 · Why this food No food-decision explanation was captured for this path.

7 · Lookup method No lookup method was captured for this path.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould log the entry"Record one RXBAR blueberry protein bar." (difficulty —)0.6s

Verdict Expected LOG — should log the entry. PASS: Logged (write-truth verified): RXBAR Blueberry Protein Bar — card not captured.

Why verdict Logged (write-truth verified): RXBAR Blueberry Protein Bar — card not captured.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.6s

1 · TTS said Logged RXBAR Blueberry Protein Bar with an estimated macro range.

2 · Card shown Logged RXBAR Blueberry Protein Bar — the macros are an estimate since this food varies by brand.

3 · App data rows written created food_log_entry: RXBAR Blueberry Protein Bar ×1 (1 bar) 210 cal · 12g P · 24g C · 8g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:21:16.309Z

6 · Why this food No food-decision explanation was captured for this path.

7 · Lookup method No lookup method was captured for this path.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould log the entry"Add a Premier Protein caramel shake." (difficulty —)1.1s

Verdict Expected LOG — should log the entry. PASS: Logged (write-truth verified): Caramel High Protein Shake — card not captured.

Why verdict Logged (write-truth verified): Caramel High Protein Shake — card not captured.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 1.1s

1 · TTS said Logged Caramel High Protein Shake — premier protein with an estimated macro range.

2 · Card shown Logged Caramel High Protein Shake — premier protein — the macros are an estimate since this food varies by brand.

3 · App data rows written created food_log_entry: Caramel High Protein Shake ×1 (1 portion (311.844 g)) 159 cal · 29.9g P · 1.9g C · 3.1g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:21:28.610Z

6 · Why this food No food-decision explanation was captured for this path.

7 · Lookup method No lookup method was captured for this path.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

FAILshould log the entry"Track one Dave’s Killer Bread thin-sliced Good Seed slice." (difficulty —)5.3s

Verdict Expected LOG — should log the entry. FAIL: OVER-ASK — asked instead of logging (no saved row).

Why verdict OVER-ASK — asked instead of logging (no saved row).

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 5.3s

1 · TTS said Was it plain, or did you have anything on it like butter or cream cheese? I did not log it yet because add-ons can change the nutrition.

2 · Card shown Was it plain, or did you have anything on it like butter or cream cheese? I did not log it yet because add-ons can change the nutrition.

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:21:46.605Z

6 · Why this food No food-decision explanation was captured for this path.

7 · Lookup method No lookup method was captured for this path.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

FAILshould log the entry"Log a Siggi’s strawberry skyr cup." (difficulty —)1.0s

Verdict Expected LOG — should log the entry. FAIL: OVER-ASK — asked instead of logging (no saved row).

Why verdict OVER-ASK — asked instead of logging (no saved row).

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 1.0s

1 · TTS said What food did you have?

2 · Card shown What food did you have?

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:22:08.990Z

6 · Why this food No food-decision explanation was captured for this path.

7 · Lookup method No lookup method was captured for this path.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

FAILshould log the entry"I had a Panera Fuji apple salad with chicken, full size." (difficulty —)2.0s

Verdict Expected LOG — should log the entry. FAIL: OVER-ASK — asked instead of logging (no saved row).

Why verdict OVER-ASK — asked instead of logging (no saved row).

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 2.0s

1 · TTS said Which full size should I use?

2 · Card shown Which full size should I use?

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:22:22.107Z

6 · Why this food No food-decision explanation was captured for this path.

7 · Lookup method No lookup method was captured for this path.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

FAILshould log the entry"Record one Celsius peach vibe can." (difficulty —)0.9s

Verdict Expected LOG — should log the entry. FAIL: OVER-ASK — asked instead of logging (no saved row).

Why verdict OVER-ASK — asked instead of logging (no saved row).

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.9s

1 · TTS said I couldn't find Record One Celsius Peach Vibe Packaged Food in my nutrition database. What exact name is on the package? I did not log it because I'd only be guessing the nutrition. If you prefer, tell me to log a generic packaged food.

2 · Card shown I couldn't find Record One Celsius Peach Vibe Packaged Food in my nutrition database. What exact name is on the package? I did not log it because I'd only be guessing the nutrition. If you prefer, tell me to log a generic packaged food.

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:22:34.199Z

6 · Why this food No food-decision explanation was captured for this path.

7 · Lookup method No lookup method was captured for this path.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

FAILshould log the entry"Add a Kodiak Cakes protein waffle, one waffle." (difficulty —)2.4s

Verdict Expected LOG — should log the entry. FAIL: WRITE-TRUTH FAIL — WRONG item/product identity "Kodiak Cakes protein waffle" — saved row does not hit product alias (rows: Birthday Cake Power Waffles With Sprinkles, Waffle, Plain, Waffle, Plain)

Why verdict WRITE-TRUTH FAIL — WRONG item/product identity "Kodiak Cakes protein waffle" — saved row does not hit product alias (rows: Birthday Cake Power Waffles With Sprinkles, Waffle, Plain, Waffle, Plain)

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 2.4s

1 · TTS said Logged Birthday Cake Power Waffles With Sprinkles — KODIAK and Waffle, Plain. Assumed Birthday Cake Power Waffles With Sprinkles — KODIAK for "a Kodiak Cakes protein waffle" (best matching option). Tell me if that is wrong. Tell me if that is not right.

2 · Card shown Logged Birthday Cake Power Waffles With Sprinkles — KODIAK and Waffle, Plain. Assumed Birthday Cake Power Waffles With Sprinkles — KODIAK for "a Kodiak Cakes protein waffle" (best matching option). Tell me if that is wrong. Tell me if that is not right.

3 · App data rows written created food_log_entry: Birthday Cake Power Waffles With Sprinkles ×1 (1 waffle) 230 cal · 10g P · 27g C · 9g F
created food_log_entry: Waffle, Plain ×1 (1 small waffle) 149 cal · 3.5g P · 16.6g C · 7.5g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:22:47.796Z

6 · Why this food No food-decision explanation was captured for this path.

7 · Lookup method No lookup method was captured for this path.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

FAILshould log the entry"Track a Subway six inch turkey on wheat with provolone." (difficulty —)3.1s

Verdict Expected LOG — should log the entry. FAIL: OVER-ASK — asked instead of logging (no saved row).

Why verdict OVER-ASK — asked instead of logging (no saved row).

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 3.1s

1 · TTS said What exact food and amount should I use for a subway six inch turkey on wheat with provolone? I did not log it yet because I could not safely finish that food log.

2 · Card shown What exact food and amount should I use for a subway six inch turkey on wheat with provolone? I did not log it yet because I could not safely finish that food log.

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:23:02.149Z

6 · Why this food No food-decision explanation was captured for this path.

7 · Lookup method No lookup method was captured for this path.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

FAILshould log the entry"Log one KIND dark chocolate nuts and sea salt bar." (difficulty —)0.5s

Verdict Expected LOG — should log the entry. FAIL: WRITE-TRUTH FAIL — WRONG item/product identity "KIND dark chocolate nuts" — saved row does not hit product alias (rows: Dark chocolate, No Cow Chocolate Sea Salt Bar)

Why verdict WRITE-TRUTH FAIL — WRONG item/product identity "KIND dark chocolate nuts" — saved row does not hit product alias (rows: Dark chocolate, No Cow Chocolate Sea Salt Bar)

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.5s

1 · TTS said Logged Dark chocolate and No Cow Chocolate Sea Salt Bar — No Cow. Assumed catalog default servings where you did not say an amount — tell me if that is not right.

2 · Card shown Logged Dark chocolate and No Cow Chocolate Sea Salt Bar — No Cow. Assumed catalog default servings where you did not say an amount — tell me if that is not right.

3 · App data rows written created food_log_entry: Dark chocolate ×1 (100 g) 550 cal · 5.1g P · 60g C · 32.2g F
created food_log_entry: No Cow Chocolate Sea Salt Bar ×1 (1 bar (60 g)) 210 cal · 20g P · 23g C · 7g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:23:13.789Z

6 · Why this food No food-decision explanation was captured for this path.

7 · Lookup method No lookup method was captured for this path.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

Why the fails happened — comprehension vs execution vs cosmetic

Handled correctly? — by expected action

Accuracy by difficulty

Clarification follow-ups — scored separately

Cosmetic only

System / infra

Latency

Filter — controls the list below

Per-utterance detail