API Drift Detective

APIs change. Versions drift. For consumers - internal or external - the question is not 'did the API change', it is 'what changed, and does it matter for me?'.

This experiment focuses on API / schema comparison and change detection: given two versions of an API, what is different, and how do you present that difference in a way that is actually useful?

• Practical API / schema comparison: surfacing meaningful changes, not just textual diffs.
• Change detection as a first-class product surface rather than a side effect of a CI script.
• Using SignalR to push change events to a UI in real time.
• Where AI fits naturally into API change reporting, and where it does not.

• An API change detection platform that compares API / schema versions.
• A UI to inspect the diff and understand the impact of changes.
• SignalR for live push of detected changes to connected clients.
• AI-assisted interpretation of detected changes as an area of exploration.

The backend does the comparison work - fetching API specs, computing the diff, and publishing change events. The frontend consumes those events and presents the diff in a readable form.

• .NET backend for comparison and change detection.
• React + TypeScript frontend for the inspection experience.
• SignalR for pushing change events to the UI.
• AI exploration layered on top of detected changes where it adds value.

• API comparison is easy to start and surprisingly hard to make genuinely useful - context is everything.
• Live push of change events changes the UX in ways a polling approach cannot.
• AI is a tool, not a strategy: it helps where the underlying diff is already solid.
• Change detection is most valuable when it tells a story about impact, not just about lines.

Active build. The next areas of focus are tightening the comparison engine, expanding the inspection UI, and evaluating where AI-assisted interpretation adds the most value.