Quick Start

This guide gets you from zero to a live browser mock in five minutes using msw-cli. Make sure you’ve completed Installation first.

With msw-cli

Open a session (starts the daemon on port 6789):

msw-cli open

Output:

[MSW] Daemon started on ws://localhost:6789
[MSW] Session: my-project

Check the browser connection:
Terminal window
```
msw-cli status
```
You should see connected: true once your app has loaded in the browser with the MSW bridge active.
Add a mock handler:
Terminal window
```
msw-cli add "http.get('/api/user', () => HttpResponse.json({ id: 1, name: 'Alice' }))"
```
The handler activates in your browser immediately — no reload needed.
Verify it’s working:

In DevTools or your app, make a request to /api/user — you’ll see the mocked response:
```
{ "id": 1, "name": "Alice" }
```

Update the mock:

msw-cli update "/api/user" -h "http.get('/api/user', () => HttpResponse.json({ id: 2, name: 'Bob' }))"

If you have msw-mcp configured in your MCP client (Cursor, Claude Code, etc.):

Scaffold your project — run the /msw-setup prompt:
```
/msw-setup
```
The AI will detect your framework, install dependencies, and create the mocks folder.
Start your dev server so the browser connects to the MSW WebSocket.
Ask your AI assistant to add a mock:

“Mock the GET /api/products endpoint to return an array of 3 products with id, name, and price.”

The AI generates the handler code and calls msw_add_handlers — the browser updates instantly.
Test an error scenario:

“Make POST /api/checkout return a 503 Service Unavailable error.”
Check what’s active:

“What MSW handlers are currently active?”

The AI calls msw_get_status and reports back.