# Kane CLI

> Terminal testing agent from TestMu AI that verifies plain-English browser and mobile flows in real Chrome and returns an evidence-backed pass or fail.

Kane CLI is a command-line testing agent from TestMu AI (formerly LambdaTest). Developers and coding agents describe a user flow in plain English, and Kane CLI runs it in a real Chrome browser or on mobile targets and returns a pass or fail with an evidence pack. It is installed via npm or brew.

## What It Is

Kane CLI brings TestMu AI's KaneAI end-to-end testing agent to the terminal. It consists of a Web CLI for browser automation and a Mobile CLI for mobile automation. Instead of writing selectors or scripts, a user writes an objective such as checking out as a guest and verifying the email and total. Kane CLI plans the steps, performs each in the browser, checks the results, and reports a verdict.

## How a Run Works

Each run takes context such as variables, secrets and saved logins, plans steps and checkpoints, then snapshots the page (DOM, accessibility tree, screenshot), acts like a user, and asserts DOM first with a vision fallback. Outputs include an evidence pack of steps, screenshots, console and network logs and the verdict. Runs upload to TestMu AI Test Manager by default, can post a PR comment through a GitHub Action, and can export flows as Playwright code. The vendor states a pass means every acceptance criterion is satisfied, with verification separate from the reasoning that navigates the flow.

## Healing and Saved Tests

A run can be saved as a test.md. Unchanged steps replay from cache without LLM cost, and a broken step is re-authored with the change recorded. The vendor describes step-heal and flow-heal tiers, and says a wrong value such as an incorrect total is reported as a bug rather than healed into a pass.

## Use with Coding Agents and CI

An --agent flag emits structured NDJSON so agents such as Claude Code, Cursor, Codex and Gemini CLI can parse results. For CI, --headless and --timeout are supported, with exit codes 0 for pass, 1 for fail, 2 for setup or auth errors and 3 for timeout. A spec-to-proof assurance loop links tests to requirements from sources like Jira, Confluence and Linear. Local mobile runs target Android emulators and iOS simulators and need an Apple Silicon Mac with Android Studio or Xcode; saved tests can also run on the TestMu AI cloud.

## Features
- Plain-English browser and mobile test authoring without selectors
- Real Chrome execution with DOM-first assertions and vision fallback
- Evidence pack with steps, screenshots, console and network logs, and verdict
- Auto-healing that re-authors broken steps and records changes
- Replay of saved test.md from cache without LLM cost
- Export of recorded flows to Playwright code
- Agent mode with structured NDJSON output
- Headless CI mode with defined exit codes
- GitHub Action PR comment with verdict and evidence link
- Requirement-linked assurance loop with coverage reporting
- Android emulator, iOS simulator and real device targets
- Saved logins, profiles and masked secrets and variables

## Integrations
Claude Code, Cursor, Codex, Gemini CLI, GitHub Copilot, GitHub Actions, Jira, Confluence, Linear, TestMu AI Test Manager, Playwright, Slack, HyperExecute

## Platforms
WINDOWS, MACOS, LINUX, ANDROID, IOS, WEB, API, VSC_EXTENSION, CLI

## Pricing
Subscription-based

## Links
- Website: https://www.testmuai.com/kane-cli/
- Documentation: https://www.testmuai.com/docs/kane-cli-introduction
- Repository: https://github.com/LambdaTest/kane-cli/
- EveryDev.ai: https://www.everydev.ai/tools/kane-cli