
API contracts often become outdated as implementations change. A documented API may say that an endpoint returns a particular field, parameter, response status, or data type while the actual implementation has already changed. Developers then discover these inconsistencies through broken clients, runtime errors, or manual investigation. API Contract Drift Detector automates this process. The project compares an expected OpenAPI contract with an actual OpenAPI contract and produces structured drift changes. It detects changes such as added or removed endpoints, added parameters, removed fields, field type changes, required/optional changes, and response-status changes. Each detected change includes its location, severity, expected value, actual value, and explanation. The system then analyzes a consumer repository to find source-code references affected by those contract changes. It supports common web-development languages including Python, JavaScript, TypeScript, JSX, and TSX, while avoiding irrelevant directories such as node_modules, dist, and virtual environments. Detected impacts are converted into a structured remediation context and remediation plan. The project also includes a Bob-oriented remediation workflow where affected files can be handled individually with scoped instructions. The workflow is designed so that an agent works only on the affected file and can report whether an action was resolved or could not be safely resolved. After remediation, the project can run impact analysis again to verify whether the targeted references have been resolved. This provides a machine-checkable verification step instead of relying only on an agent saying that a fix was completed. A React dashboard provides the user interface. Users can select expected and actual contracts, run analysis, inspect detected drift, view affected files and lines, and review proposed remediation actions. A FastAPI backend exposes the analysis and remediation workflow to the frontend while keeping the deterministic drift-engine logic separate from the UI and agent layer. The target users are developers and teams working with APIs who need to detect contract inconsistencies before they cause downstream problems. The main idea is to connect contract comparison → source-code impact analysis → remediation planning → verification into one workflow rather than treating API documentation and application code as separate systems.
27 Sep 2026