Star on GitHubOur Story
Users kept telling me the same thing: API testing is still a mess.
In my last company (Preflight), we built workflows around UI automation. It worked, but it had the problems you'd expect: it was premature, black-box, and not developer-friendly. UI-driven workflows are also inherently less deterministic.
So we rebuilt the idea from the right layer: the API layer.
DevTools takes real traffic (HAR files) and turns it into deterministic API workflows you can run locally and in CI. The workflows export to YAML so they're readable, diffable, and code-reviewable in Git. And because it's API-first, it's generally more secure and less flaky than UI automation for the same job.
The real unlock is making workflows practical at scale: auto variable mapping, composable flows, and a "powerful YAML" that can live in CI for any API. Long-term, that same workflow format becomes a platform: combine flows, expose endpoints, trigger integrations, and let AI generate or modify workflows safely because the output is structured and reviewable.
That's the goal: workflows as code, living in CI — not trapped in a dashboard.
If you want the full backstory, read Why we built this.
