releasetwin 0.4.1

ReleaseTwin CLI

Release-proof testing for integration-heavy, feature-flagged systems. Compose HTTP and UI journeys, run them from your own machine or CI, and prove a fix works by running the same case known-bad and known-good ("flag proof"). Test data never leaves your infrastructure.

Look first: a real case file, no account. Then install this CLI.

releasetwin.com has the pitch, a live demo, and the self-serve dashboard sign-up.

Install

dotnet tool install --global releasetwin

Needs the .NET runtime. If you'd rather not install .NET, use the container image instead — see docs/install.md.

Run

releasetwin ./cases

The CLI exits non-zero on any case failure, so it drops into a pipeline as a required check with no extra wiring. Fixtures resolve from a fixtures/ directory that is a sibling of the cases directory. ${ENV_VAR} references in case files are interpolated from the process environment. This is identical to the container image's behavior.

UI-journey cases drive a real browser via Playwright — run playwright install once. HTTP and flag-proof cases need nothing extra.

Already have a test suite?

If your suite already writes JUnit XML — pytest --junitxml, jest-junit, Playwright's junit reporter — you can put its results on the dashboard without authoring a case file:

RELEASETWIN_API_TOKEN=... releasetwin upload-junit junit-results.xml

Add --release <label> to group the upload under a release. The file is uploaded exactly as written; the platform parses it, so any dialect your runner emits is its call to accept, and it reports how many test cases were recorded.

Imported results give you run history, trends, and release roll-up. Authoring a case is what proves a flag.

More

• Quickstart
• Running in CI
• All install paths

Licensed AGPL-3.0. Independent adapters that link the AdapterSdk may carry any license — see LICENSE.EXCEPTIONS.