Paste your broken API logs, get back what went wrong and how to fix it. Multiple AI agents argue about the diagnosis so you don't have to.
LangGraphGroqFastAPIPythonNext.jsTypeScript
// why this exists
Debugging a broken API integration is usually one person grepping logs alone and guessing. SentralQ runs six specialized LangGraph agents instead: one reads the request, one reads the response, one only thinks about auth, one only validates schemas. A supervisor merges their findings into a single root cause and a concrete fix.
6s avg inference, 30+ reasoning steps
// how it's built
Debug Request Intake
- FastAPI POST /debug and /test-request
- Streaming agent progress over the same request
- CLI entrypoint for local debugging
Six-Agent LangGraph Pipeline
- Supervisor routes to Request Analyzer, Response Parser, Auth Troubleshooter, Schema Validator
- Groq-hosted Llama 3.3 70B does the reasoning
- Tool calls: decode_jwt_token, validate_json_schema, parse_openapi_spec, test_oauth_flow
Solution Generator
- Merges all four analyses into a single root cause
- Returns a concrete fix, not just a diagnosis
- Response: root_cause + solution + per-agent analysis_results