Overview

Kiwami runs your test suite — Playwright, Robot Framework, Cypress, or a custom container image — against your own repository, without you having to maintain the runner infrastructure yourself.

The shape of it:

1. Connect a GitHub repository. Kiwami clones it with a scoped GitHub App installation token, never a personal access token.
2. Create a project. Point it at a repo (or pick one from the GitHub connection) and Kiwami bakes a reproducible environment image from your own dependency lockfile.
3. Autodetection finds your suites. Kiwami looks at the repo's own structure — config files, test directories, lockfiles — and proposes the framework and test suites it found, for you to confirm or adjust.
4. Runs execute in isolated containers. Each run gets its own container, built from the project's baked image, with structured results (pass/fail per test, not just a log) parsed back out when the framework supports it.
5. A push to your default branch triggers a rebuild and a run automatically, and reports the result back to GitHub as a commit status — the same check row you'd expect from any CI system.

Billing is seat-plus-usage: every plan includes a monthly run-minute allowance, with overage billed per minute past that. See Plans & billing for the actual numbers.

Next: Quickstart.