Skip to main content
This guide takes you from a fresh install to a passing run in under five minutes.

​
Step 1: Install

Once kane-cli --version prints a version number, you are ready. See Installation for platform requirements and troubleshooting.

​
Step 2: Authenticate

Your default browser opens a consent page. Sign in (or confirm, if you are already signed in) and authorise the CLI. The page then tells you it is safe to close the tab. Control returns to your terminal, and your credentials are stored under ~/.testmuai/kaneai/profiles/.
Verify authentication at any time:
If you prefer username and access key, or need to manage multiple accounts, see Authentication.

​
Step 3: Open the TUI

Run kane-cli --tui:
You see a short startup animation: a small VR-helmet image on the left, with the words KANE CLI building up next to it in block letters. A grey footer below shows which profile you are logged in as. After the animation, you land at a chat prompt. A status bar at the bottom shows your model, session ID, environment, and run count.
The exact characters are decorative. What matters is that you reach the chat prompt at the bottom. Anything you type there is treated as a test objective, except lines that begin with /, which are slash commands such as /help, /config, /cancel, and /exit.

​
Step 4: Run Your First Test

At the chat prompt, type a natural-language objective and press Enter:
Kane CLI launches Chrome, passes the objective to the agent, and streams progress as the run goes. A step tree fills in as the agent thinks, observes the page, and acts: clicking, typing, scrolling, and verifying. When the agent finishes, a result block shows whether the objective passed or failed and how many steps it took. To stop a run, press Ctrl+C once. Pressing it twice exits Kane CLI.

​
Step 5: Or Run from the Command Line

For scripting and CI, skip the TUI and use the run subcommand:
In CLI mode, Kane CLI streams progress to stderr, prints the final result as a single JSON object on stdout, and exits with a status code that reflects the outcome: Capture results in a shell script:
By default, each run starts on the KaneAI playground site, and the agent navigates from there based on your objective.

​
What Happens at the End of a Run

When the session ends, Kane CLI uploads the run to Test Manager and prints a share link. It also saves the run as a sealed evidence pack: screenshots, logs, and results in one file. When Kane CLI offers to open it, you can view the pack in the browser viewer. For details on the upload, share links, and the run mode toggle, see Test Manager Integration. To change settings like window size, Chrome profile, or the active project and folder, see Configuration.
Run kane-cli --tui to open the interactive TUI: a full terminal UI where you can type objectives, see step-by-step progress, and chain multiple tests in one session with persistent browser state.

​
Next Steps