API tests from specs — and risks from a live cURL.
Upload Postman or OpenAPI/Swagger to generate NLP coverage. Or paste one working cURL to discover validation, auth, boundary, and protocol failures — run in cloud or on your local agent, then promote failures into regression and Engineering Memory.
Real product UI — API testing
Screenshots from the TestNeo app: NLP API cases, coverage graph, flows, and run analysis.
Postman · Swagger · Impact — interactive demo
The same import panel from the homepage feature theater: import a collection or OpenAPI file, see generated coverage, then watch schema drift flag linked tests.
- Postman JSON → NLP suite with auth and chaining preserved
- OpenAPI YAML/JSON → happy + negative + boundary + auth
- Schema changes trigger impact analysis on linked tests
- No spec yet? Paste a cURL for Risk Discovery
- Export the same suite to k6 for load
API Risk Discovery
Have a working request but no suite yet? Paste one cURL. TestNeo expands it into deterministic validation, boundary, auth, schema, and protocol variants — no LLM inventing attacks — then scores API Confidence. Failures promote to NLP regression and Engineering Memory so the next PR can WARN.
- Discover → Run (cloud or local agent) → Promote & remember
- Protocol signals: HTML-from-JSON, soft method handling, security-header hygiene
- Export Confidence Brief or Postman collection
- Complements Swagger/Postman generation — not a replacement
From a spec
Postman / OpenAPI → broad NLP coverage and schema impact.
From a live cURL
Risk Discovery → break assumptions on one endpoint, then feed RI.
# Working request — FakeStore demo curl -X POST 'https://fakestoreapi.com/products' \ -H 'Content-Type: application/json' \ -d '{"title":"Demo Jacket","price":55.99}'
Toy APIs that accept garbage correctly score near zero — that’s the signal.
Discover → Verify → Remember → Decide — scanners stop at findings; TestNeo turns them into release evidence.