Start with one reproducible failure
A screenshot of a bad answer is rarely enough to identify the cause. Keep the input, the source documents and the tool response together. Write down the expected behavior in a sentence someone else can judge.…
DIAGNOSE · TEST · IMPROVE
AI Agent Support
Turn a failing AI workflow into a practical debugging checklist and a clear support ticket.
✦ Free to use↗ Export your result○ No account required
01
Make it yours
02
Your result
Ready when you areA little input.
A useful starting point.
An agent cites a policy that is absent from the retrieved documents.
Example · customize the fields to beginEdit the text before you export
Rule-based troubleshooting paths selected by failure category; all inputs are processed in your browser.
Find the failing layer
Agent failures often start before the final model response. Missing documents belong to retrieval, expired tokens belong to the integration, and repeated write actions belong to orchestration. Test each layer…
Build a regression set
Keep a short collection of real failures, normal requests and deliberately incomplete inputs. After each change, run the entire set and record what improved and what regressed. A fluent answer is not evidence…
GOOD TO KNOW
A few useful answers.
Does this connect to my agent?
No. It builds a symptom-based troubleshooting plan locally. It does not access logs or change your systems.
Should I paste API keys or full logs?
No. Use a redacted description. Secrets are not needed to create the checklist.
Is the diagnosis certain?
No. It suggests tests to distinguish likely causes; it does not observe your runtime.
Can I send the result to a developer?
Yes. Copy or download the brief, then add a minimal reproducible example and sanitized error details.