Midscene.js alternative for web UI testing
Kane CLI is the Midscene.js alternative for web teams. Vision-grounded natural-language tests with a managed model, autoheal when the UI changes, a verified pass or fail with shareable video, and clean CI exit codes. Free to install.
or read the documentation
Why web teams move from Midscene.js to Kane CLI
Midscene.js is vision-driven and natural-language, and it can plug into Playwright, Puppeteer, or Vitest, which is great for adding AI steps to an existing suite across web, mobile, and desktop. For web QA, that flexibility still leaves you wiring a model, writing aiTap and aiAssert in code, and building the verdict and CI reporting yourself.
Kane CLI is browser-viewport disciplined and made for the web QA job: a natural-language objective drives a real Chrome, it is vision-grounded so it handles canvas and shadow DOM, and it returns a deterministic pass or fail with video and CI exit codes. The model is managed, so there is nothing to choose or self-host.
Need mobile or desktop too? Kane CLI stays focused on the browser and points native needs to the broader TestMu AI platform, so the web pass or fail stays anchored to real user actions.
Kane CLI vs Midscene.js
Both are vision-grounded and natural-language. Kane CLI adds a deterministic verdict contract for web QA.
| Capability | Kane CLI (TestMu AI) | Midscene.js |
|---|---|---|
| Element grounding | Vision-based waiting and matching | Vision-driven, screenshot grounding |
| Natural-language actions | Objectives plus built-in assertions | aiTap, aiAssert, aiQuery in code |
| Setup | Standalone CLI, Playwright export when needed | Add to Playwright, Puppeteer, or Vitest |
| Model | Managed model, nothing to configure | Bring or self-host a model |
| Coding-agent skills | Skills for Claude Code, Codex, Gemini | Skills and CLIs |
| Verified pass or fail | Deterministic verdict, NDJSON, CI exit codes | Assertions, no verdict contract |
| Native Playwright export | One command | |
| Scope | Browser-focused, native routes to the platform | Web, Android, iOS, desktop |
What you get with Kane CLI
Everything Midscene.js leaves to your suite, handled by the tool.
No code, no keys
Skip the SDK wiring and the model and API key setup. Install one CLI, describe the journey in natural language, and Kane CLI drives a real browser with a managed model.
Resilient by default
When the frontend changes, Kane CLI adapts on its own, pushing through up to 50 steps until the full journey is verified instead of breaking on a moved element.
Verified pass or fail
Built-in assertions check each step, not just the final screen, and flag the exact point of failure. A real result, not raw agent output to parse.
Replay and caching
Validated flows replay from a cached test.md with no repeat LLM cost, so re-running the suite in CI does not keep burning tokens.
Cross-browser cloud and CI
Scale to cross-browser and cross-OS runs on the TestMu AI grid with one flag, and gate pipelines on clean exit codes from your terminal.
Own the Playwright code
Export any validated flow to native Playwright with one command, then modify and own it. The convenience never locks you in.
Build up confidence locally
Start in your terminal
Validate on the cloud
Release with confidence
From prototype to production, on one engine
Kane CLI and KaneAI share the same automation engine and dashboard.
Prototype to production
Open-source agents and SDKs are great for prototyping. Kane CLI is the production-grade, test-focused path: assertions, caching, evidence, and CI built in.
Built for the agent loop
Generate code, then verify it in a real browser before opening a PR. Kane CLI closes the gap between what an agent writes and what works.
Evidence you can share
Every run produces a persistent video, step trace, and replay link to drop into a PR or bug report, backed by a dashboard your team can browse.
Replace Midscene.js in three steps
Install
Run npm install -g @testmuai/kane-cli, then sign in with your TestMu AI account. No project to scaffold and no API keys to manage.
Describe the flow
Tell Kane CLI what to do in natural language, or point your AI coding agent at the agent guide and let it drive the browser.
Run and verify
Kane CLI drives a real browser, verifies each step, and returns a pass or fail with video proof. Re-runs replay from cache with no repeat LLM cost.
Get Started With Kane CLI
🎉 Launch offer: Bonus credits for the first 3 months on paid plans
Choose the right plan for you
Local test authoring via CLI
Auto-heal & vision
View test cases on UI
Test Manager
Free
$0
/month
200 Credits
Resets in every
30 days
Starter
$19
/month
2000 Credits
Launch: 4,000 Credits (+100%)
Bonus for first 3 months
Pro
$99
/month
10,000 Credits
Launch: 15,000 Credits (+50%)
Bonus for first 3 months
Enterprise
Get access to solutions built on Enterprise-Grade Security, Privacy, and Compliances.
Need more credits?
Got a bigger use case in mind?
Let’s talk
Choose the right plan for you
Free
$0
/month
200 Credits
Resets in every
30 days
Starter
$19
/month
2000 Credits
Launch: 4,000 Credits (+100%)
Bonus for first 3 months
Pro
$99
/month
10,000 Credits
Launch: 15,000 Credits (+50%)
Bonus for first 3 months
Enterprise
Get access to solutions built on Enterprise-Grade Security, Privacy, and Compliances.
Need more credits?
Got a bigger use case in mind?
Let’s talk
Get the technical rundown
Documentation
Everything you need to install, configure, and run Kane CLI in under 2 minutes.
Frequently asked questions
Kane CLI by TestMu AI is the leading Midscene.js alternative for web UI testing. Midscene.js is a vision-driven, natural-language automation tool you add to Playwright, Puppeteer, or Vitest, with a model you bring or self-host. Kane CLI takes a natural-language objective from any terminal, drives a real browser with a managed model, is vision-grounded for canvas and shadow DOM, and returns a verified pass or fail with shareable video. No model to choose and no suite to wire.
Midscene.js is an open-source, vision-driven UI automation and testing tool from the ByteDance Web Infra team. It uses screenshot-based element grounding and natural-language APIs such as aiTap and aiAssert, works across web, Android, iOS, and desktop, and integrates with Playwright, Puppeteer, and Vitest. The model can be brought or self-hosted.
Kane CLI is a standalone CLI focused on the browser, with a managed model so there is nothing to choose or self-host. It returns a deterministic pass or fail with NDJSON and clean CI exit codes, self-heals with confidence-scored matching, replays from a cached test.md with no repeat LLM cost, and exports to native Playwright with one command.
Kane CLI is browser-focused by design, which keeps its pass or fail anchored to real user actions in a real browser. For native mobile and desktop testing, the broader TestMu AI platform covers those surfaces, so you get a web verdict from Kane CLI and native coverage from the platform.
No. Kane CLI ships a managed model, so you install it and go. Validated flows also replay from a cached test.md with no repeat LLM cost, so frequent CI re-runs do not keep burning tokens.
Teach your agent the right skills
Point your AI coding agent at the Kane CLI guide and it will install, authenticate, and run verified browser flows on its own, no SDK or API keys required.
Point your agent to: testmuai.com/kane-cli/agents.md