Kane CLI By TestMu AI

About Kane CLI By TestMu AI

Kane CLI is the terminal-native AI testing tool from TestMu AI (formerly
LambdaTest). Describe what should happen in plain English; Kane CLI runs it in
a real Chrome browser, verifies every step, and returns pass or fail with a
shareable evidence link. No selectors, no test scaffolding to maintain.

THE GAP IT FILLS

AI coding agents like Claude Code, Cursor and Codex CLI can build entire
features from a prompt, but they cannot verify that what they built actually
works in a browser. General-purpose browser agents drive a browser but have no
concept of a test — no assertions, no pass/fail, no evidence trail. Playwright,
Selenium and Cypress are powerful but need a dev team, setup, and upkeep every
time the UI shifts.

Kane CLI sits between them: natural language in, real browser execution,
deterministic pass or fail with proof.

GETTING STARTED

npm install -g @testmuai/kane-cli
kane-cli login
kane-cli run "verify checkout works end-to-end"

Requires Node.js 18+. Also on Homebrew. Runs on macOS, Linux and Windows.

WHY RUNS HOLD UP

Autoheal absorbs cosmetic UI changes that would break a hand-written selector.
Vision-based dynamic waiting detects loaders and animations on screen rather
than relying on network events, so canvas, shadow DOM and stubborn framework
components are handled. Kane CLI only takes actions a real human could take —
it will not inject JavaScript to bypass a validation and force a green result.
Runs push through up to 50 steps per flow rather than returning halfway.

THREE MODES

Interactive TUI for development, with the browser staying open between runs.
Headless CLI for shell scripts and CI, with standard exit codes and JSON to
stdout. Agent Mode for AI coding agents, emitting structured NDJSON — one JSON
object per line — so agents can parse results programmatically.

FOR AI AGENTS

One command installs Kane CLI as a native skill for Claude Code, Codex CLI and
Gemini CLI:

npx @testmuai/kane-cli-skill

Agent instructions: testmuai.com/kane-cli/agents.md. An Ask tool keeps a human
in the loop for OTPs and CAPTCHAs without abandoning the run.

NO LOCK-IN

Every flow exports to native Playwright code when you want to own it, and
existing Playwright or Selenium suites can be migrated in. Test.md captures any
session as replayable markdown with imports, variables and replay. Runs sync to
Test Manager for history and step-by-step logs.

Bui