Project profile
Scope, role and technology
Visual evidence
Test-run screenshots, structured XLSX report extracts, failure examples, regression cases, and correction/rerun evidence from the production API audit workflow.
Why the API needed its own auditor
BAYNAO’s product API sits between business data and AI/customer workflows. Small API regressions can propagate into wrong search results, missing alternatives, incomplete product information, or confusing customer responses.
- Built automated tests instead of relying only on manual endpoint checks.
- Tested real selected products and query variants against expected behavior.
- Used findings to iteratively correct API logic and rerun regression cases.
What the test system checks
- Missing or malformed fields.
- Incorrect or inconsistent product relations.
- Unexpected search/routing behavior and ambiguous query collisions.
- Product-response consistency across different request forms.
- Regression checks after code corrections.
Persistent audit state
A local audit-state database preserves selected entities, execution status, failure evidence, and progress across longer runs so testing can continue without losing what has already been evaluated.
Evidence and correction reports
Results are exported into structured reports that separate pass/fail or issue categories, preserve response evidence, and make recurring failure classes easier to debug.
Architecture
Engineering decisions
- Convert discovered edge cases into repeatable regression cases.
- Keep test state and evidence outside the production API.
- Do not accept a fix until the relevant regression case passes again.