Skip to main content
FeatureFactory
Free tool · runs in your browser

Acceptance Criteria Grader

Paste your acceptance criteria and get a testability score, per-line flags, and specific fixes - before they reach a sprint.

|

Bullets and numbering are stripped automatically. Heuristics only - an editable lint signal, not a verdict.

D
52/100
Overall testability
Criteria
5
With G/W/T
2
Testable
3
Ambiguous
2

1.Given a logged-out user, when they submit valid credentials, then they are redirected to /dashboard within 2 seconds.

100
Given/When/Then structureObservable verbHas a concrete value

Looks solid - nothing to fix.

2.The search results should load fast and feel easy to use.

0
No Given/When/ThenNo observable verbNothing measurable to assertAmbiguous: fast, easy
  • →Restructure as "Given <context>, when <action>, then <observable result>".
  • →State the concrete system behaviour (returns, displays, redirects, sends…).
  • →Add a checkable value: a number, status code, path, or exact string.
  • →Replace vague words (fast, easy) with a specific threshold.

3.When the export button is clicked, then a CSV file containing all 12 columns is downloaded.

100
Given/When/Then structureObservable verbHas a concrete value

Looks solid - nothing to fix.

4.The API returns a 429 status code after more than 100 requests per minute from a single IP.

60
No Given/When/ThenObservable verbHas a concrete value
  • →Restructure as "Given <context>, when <action>, then <observable result>".

5.Users can somehow filter the list.

0
No Given/When/ThenNo observable verbNothing measurable to assertAmbiguous: somehow
  • →Restructure as "Given <context>, when <action>, then <observable result>".
  • →State the concrete system behaviour (returns, displays, redirects, sends…).
  • →Add a checkable value: a number, status code, path, or exact string.
  • →Replace vague words (somehow) with a specific threshold.

What makes acceptance criteria testable?

Acceptance criteria fail in review for boring, repeatable reasons: they describe a feeling ("fast", "easy") instead of an observable behaviour, they skip the context that makes a scenario reproducible, or they name no value a test can assert against. This grader checks each line for a Given/When/Then structure, an observable verb, a concrete value, and the absence of ambiguity words - the same lint you would apply by hand, only faster.

Good criteria are the contract between what you plan and what you ship. If you want the LLM in your pipeline to generate strong scenarios in the first place, pair this with our acceptance criteria generator and the product plan generator. For the reasoning behind requirement quality, read the guide on how to write a product plan.

Testable criteria are also the front door to measurement. When outcomes are stated as checkable behaviours, you can trace them all the way through to shipped code with measure and see the rest of the workflow on the product tour.

Frequently asked

Each line is scored out of 100 across four heuristics: a Given/When/Then (or When/Then) structure, an observable verb, a concrete testable value (number, status code, path, quoted string), and the absence of ambiguity words. The overall score is the average across all non-empty lines. It is a lint-style signal, not a substitute for review.

Measure what you ship, from criteria to code

FeatureFactory helps engineering teams plan, build, and measure AI web development - with criteria that actually hold up in review.

Get early access