Skip to main content
This page is a short procedure for a developer or a coding agent. A coding agent can read this page as plain Markdown at https://docs.qawolf.com/onboard.md. The full documentation index is at https://docs.qawolf.com/llms.txt.

​
Before you begin

Make sure that you have these items:
  • A QA Wolf workspace. Your team administrator must invite you.
  • An API key. Create one in the workspace under Workspace settings → API keys.
  • Node.js 22.12 or later, if you install the CLI with npm.

​
Step 1: Install the CLI

Install the CLI with npm:
If you do not want Node.js, download a standalone binary from GitHub Releases. Builds exist for Linux (x64 and arm64), macOS (Apple Silicon and Intel), and Windows x64. Make sure that the CLI is on your PATH:

​
Step 2: Connect the CLI to your workspace

If you work on a local machine, log in with your API key:
The CLI asks for the API key and stores it in the system keychain. If you work in CI or in an agent sandbox, set the QAWOLF_API_KEY environment variable instead:
Make sure that the connection works:
The output shows the team name and the team ID. Read more: Authenticate the QA Wolf CLI →

​
Step 3: Find your environment ID

Each environment has an ID. You can find it in the QA Wolf dashboard under Settings → Environments. The commands below use <env_id> for this value. Read more: Pull an environment →

​
Step 4: Pull your flows and install the runtime

Pull the flows of the environment to the local cache:
The CLI stores the flows in .qawolf/<env_id>/. Then install the browsers and the other runtime dependencies that these flows need:
Read more: Install runtime dependencies →

​
Step 5: Run your flows locally

List the flows in the local cache, then run them:
The exit code is 0 when all flows pass and 1 when one or more flows fail. Read more: Run flows locally →

​
Step 6: Start a run on the platform

To run the environment on QA Wolf infrastructure instead of your machine:
This command requires authentication from Step 2. The flows do not run on your machine. Read more: CLI command reference →

​
Step 7: Install the agent skill (optional)

The CLI repository ships a qawolf-cli skill that obeys the Agent Skills specification. Coding agents such as Claude Code, Codex, Cursor, and Gemini CLI can use it.

​
If a step fails

Run the diagnostic command. It reports the CLI version, the Node.js version, the API key, the connectivity to QA Wolf, and the installed browsers.
Read more: Diagnose problems →

​
Next steps

Quick start

Map your application and generate your first flow with AI.

Key concepts

Learn how workspaces, environments, flows, and runs are organized.

CLI reference

Every command, flag, environment variable, and exit code.
Last modified on September 4, 2026