Skip to content

Getting started

Argos ships as a desktop app and a single-binary CLI. Both read the same workspace folder, so you can hack in the UI and re-run the same requests in CI without exporting anything.

Install the desktop app

Pre-built bundles are attached to every GitHub release.

  • macOS: Argos-<version>.dmg (Apple Silicon and Intel).
  • Windows: Argos-<version>-setup.exe.
  • Linux: .deb / .rpm for the common distros, AppImage as a fallback.

The desktop app bundles the CLI binary so the two can’t drift. On first launch the welcome screen offers Open workspace folder or Create new workspace.

Install the CLI on a CI runner

The CLI is a static binary — drop it on the runner before invoking it:

Terminal window
curl -fsSL \
"https://github.com/argos-app/argos/releases/latest/download/argos-x86_64-linux.tar.gz" \
| tar -xz
chmod +x ./argos
./argos --help

See CI integration for ready-made GitHub Actions and GitLab CI workflows.

Create your first request

In the desktop app:

  1. Open workspace folder → pick an empty folder.
  2. Click + in the sidebar header → New REST request.
  3. Enter https://httpbin.org/get in the URL bar.
  4. Press +Enter (or click Send).

The response pane shows status, headers, body, and timing. Argos auto-saves the request to <workspace>/collections/<slug>.argos.yaml — open it in your editor of choice if you want to see the on-disk shape.

Add a test

In the Scripts tab of the request editor:

bru.test('status 200', () => {
bru.expect(bru.res.status).toBe(200);
});

Send the request again. The Tests sub-tab in the response pane shows ✓ status 200. The same script runs in CI when you use argos run.

Run a collection from the command line

From the workspace root:

Terminal window
argos run ./collections

Argos walks the tree, runs every request in order, prints a Mocha-like summary, and exits non-zero if anything failed. Add a reporter to capture a structured artifact for CI:

Terminal window
argos run ./collections \
--reporter junit=report.xml \
--reporter json=report.json

Next steps