CheckcraftGenerate test drafts from documentation and screenshots
Checkcraft is a desktop app for checking APIs and logs without code, with a free edition. Download, no sign-up →
A copied table can contain the field, type, requirement and expected value. A screenshot can contain the same information, but recognition errors make an extra review necessary.
You will check the recognized material, review the proposed steps and prove the resulting test with a deliberate error.
Prefer the original text when available
Copy documentation with its headings and tables or attach a structured example. Describe which operation and result you want tested. For an image, use a readable crop with complete column headings and values rather than a scaled-down screen containing several unrelated sections.
Inspect recognition before generation
Open the recognized material and compare paths, punctuation, numbers and required-field markers with the original. OCR can confuse similar characters or flatten a table. Fix the input before asking the model to reason about it. A recognition percentage is not a guarantee that each important value is correct.
Separate requirements from suggestions
Check that generated expected values come from your input. Review the selected step types and target location, then add the drafts. If the source lacks an expected result, state it explicitly. For large objects, define which paths, fields and array elements are in scope instead of relying on an unspecified request to check everything.
Repeat the check with ready-made files
Download the Checkcraft test case and two inputs. One input is valid, and the other contains a deliberate error. These synthetic examples are available without registration.
Run the example in Checkcraft
Create a collection, a check and a test case using the workspace tree context menu. Click Import in the test case editor and choose the downloaded test case. Open the added case, copy the valid JSON, click Paste data for analysis, then Run test case. Open Result to inspect the report.
Repeat with the broken JSON without changing the rules. The first run should pass and the second should fail. If you see a different outcome, check the selected data source, imported steps and report.
The example files were executed against the application source on 14 September 2026. Reference reports in English are available for the passing run and failing run. Sample names and literal data values do not change when you switch the interface language.
Inspect the draft from the screenshot
The actual AI-created draft checks orderId, status and amount in an HTTP response body. It needs an HTTP run with a matching response. The simpler pasted-data example above is a separate way to learn how passing and failing checks work.
Review AI-generated tests before reuse
A model can invent a requirement or omit a condition. Compare its assertions with the documentation and execute them on passing and deliberately failing data. Local processing depends on the configured AI provider and endpoint. Test generation alone does not imply comprehensive log analysis or model training on your project.
See the result on a small example first
Start with the sample, inspect the failure and then try a sanitized example from your own API. Free is available for Windows and Linux with no time limit.
Download FreeDownload and first runNo sign-up or email needed, the archive is ready in a couple of minutes. See the edition limits.