Summary checker
Faithful, complete and free of added opinions: the three ways summaries go wrong.
Checks an AI summary against its document: nothing invented, the key point kept, no critical caveat dropped, and no opinions the author never gave.
Try it live
Real calls to Jev. Pick an example or paste your own.
Pick an example or paste your own input, then run it.
Under the hood
Every question and rule, in the open. Fork it to change any of them.
What it asks
Reads document and summary, then asks 5 questions in one call.
- Q1FaithfulYes / no
Is everything the
summarysays supported by thedocument? - Q2Keeps the key pointYes / no
Does the
summaryinclude the single most important point of thedocument? - Q3Drops something criticalYes / no
Does the
summaryleave out something a reader needs to avoid a wrong conclusion (a caveat, a condition, or a number that changes the meaning)? - Q4Adds opinionsYes / no
Does the
summaryadd opinions, judgments or recommendations that thedocumentdoesn't make? - Q5ConcisionScore
How concise is the
summary?- 0Padded: repeats itself or wanders
- 1Could be tighter
- 2Tight
How it decides
Rules run in code, top to bottom. The first one that matches gives the verdict.
- 1If Faithful < 0.5Unfaithful
- 2If Drops something critical ≥ 0.6Needs editsreview
- 3If Adds opinions ≥ 0.6Needs editsreview
- 4If Keeps the key point < 0.5Needs editsreview
- ·DefaultGood summary
Call it from your code
Every template runs as-is with an API key, no setup: POST /api/v1/judges/summary-checker/run.
curl -X POST https://tryevally.com/api/v1/judges/summary-checker/run \ -H "Authorization: Bearer $EVALLY_API_KEY" \ -H "Content-Type: application/json" \ -d '{ "input": { "document": "The new drug reduced symptoms by 40% in adults aged 18-45. It showed no benefit in patients over 65, and 12% of participants reported headaches. Further trials are needed before approval.", "summary": "A new drug cuts symptoms by 40% and is ready to help patients of all ages." }}'More in AI output quality
- AI output quality
RAG answer grader
Is the answer grounded in the retrieved source, and does it answer the question?
PassNeeds reviewFail5 questions - AI output quality
Hallucination detector
Catch invented numbers, names and claims before your users do, with the sentence quoted.
GroundedSuspectHallucinated4 questions - AI output quality
Citation checker
Does the cited passage actually support the claim it's attached to?
SupportedPartly supportedUnsupported+12 questions - AI output quality
Support reply grader
Check an AI support reply against the customer's message and your policy before it's sent.
Ready to sendEdit firstEscalate+15 questions - AI output quality
Claim vs evidence
Does the evidence support the claim, contradict it, or not settle it? With the sentence that decides.
SupportedNot settledContradicted3 questions