Adds revolutionary features for MCP client identification and browser automation:
MCP Client Debug System:
- Floating pill toolbar with client identification and session info
- Theme system with 5 built-in themes (minimal, corporate, hacker, glass, high-contrast)
- Custom theme creation API with CSS variable overrides
- Cross-site validation ensuring toolbar persists across navigation
- Session-based injection with persistence across page loads
Voice Collaboration (Prototype):
- Web Speech API integration for conversational browser automation
- Bidirectional voice communication between AI and user
- Real-time voice guidance during automation tasks
- Documented architecture and future development roadmap
Code Injection Enhancements:
- Model collaboration API for notify, prompt, and inspector functions
- Auto-injection and persistence options
- Toolbar integration with code injection system
Documentation:
- Comprehensive technical achievement documentation
- Voice collaboration architecture and implementation guide
- Theme system integration documentation
- Tool annotation templates for consistency
This represents a major advancement in browser automation UX, enabling
unprecedented visibility and interaction patterns for MCP clients.
Implements comprehensive solution for browser_click and other interactive tools
returning massive responses (37K+ tokens) due to full page snapshots.
Features implemented:
1. **Snapshot size limits** (--max-snapshot-tokens, default 10k)
- Automatically truncates large snapshots with helpful messages
- Preserves essential info (URL, title, errors) when truncating
- Shows exact token counts and configuration suggestions
2. **Optional snapshots** (--no-snapshots)
- Disables automatic snapshots after interactive operations
- browser_snapshot tool always works for explicit snapshots
- Maintains backward compatibility (snapshots enabled by default)
3. **Differential snapshots** (--differential-snapshots)
- Shows only changes since last snapshot instead of full page
- Tracks URL, title, DOM structure, and console activity
- Significantly reduces token usage for incremental operations
4. **Enhanced tool descriptions**
- All interactive tools now document snapshot behavior
- Clear guidance on when snapshots are included/excluded
- Helpful suggestions for users experiencing token limits
Configuration options:
- CLI: --no-snapshots, --max-snapshot-tokens N, --differential-snapshots
- ENV: PLAYWRIGHT_MCP_INCLUDE_SNAPSHOTS, PLAYWRIGHT_MCP_MAX_SNAPSHOT_TOKENS, etc.
- Config file: includeSnapshots, maxSnapshotTokens, differentialSnapshots
Fixes token overflow errors while providing users full control over
snapshot behavior and response sizes.
Co-Authored-By: Claude <noreply@anthropic.com>
Add comprehensive artifact storage system with session-specific directories:
- Add --artifact-dir CLI option and PLAYWRIGHT_MCP_ARTIFACT_DIR env var
- Create ArtifactManager class for session-specific artifact organization
- Implement ArtifactManagerRegistry for multi-session support
- Add tool call logging with JSON persistence in tool-calls.json
- Update screenshot, video, and PDF tools to use centralized storage
- Add browser_configure_artifacts tool for per-session control
- Support dynamic enable/disable without server restart
- Maintain backward compatibility when artifact storage not configured
Directory structure: {artifactDir}/{sessionId}/[artifacts, videos/, tool-calls.json]
🤖 Generated with [Claude Code](https://claude.ai/code)
Co-Authored-By: Claude <noreply@anthropic.com>
- Add browser_configure tool to change headless/headed mode, viewport, and user agent during session
- Fix STDIO entry point by preventing stdin close handlers in STDIO mode
- Fix headed mode default behavior when DISPLAY is available on Linux
- Add dynamic browser configuration update mechanism in Context class
🤖 Generated with [Claude Code](https://claude.ai/code)
Co-Authored-By: Claude <noreply@anthropic.com>
* HTTPS server launched and the relay server is created before MCP
server. This way we can pass CDP endpoint to its constructor.
* MCP HTTP transport is added to precreated HTTP server.
* A bunch of renames to fix style issues.
Instructions:
1. `git clone https://github.com/mxschmitt/playwright-mcp && git
checkout extension-drafft`
2. `npm ci && npm run build`
3. `chrome://extensions` in your normal Chrome, "load unpacked" and
select the extension folder.
4. `node cli.js --port=4242 --extension` - The URL it prints at the end
you can put into the extension popup.
5.
Put either this into Claude Desktop (it does not support SSE yet hence
wrapping it or just put the URL into Cursor/VSCode)
```json
{
"mcpServers": {
"playwright": {
"command": "bash",
"args": [
"-c",
"source $HOME/.nvm/nvm.sh && nvm use --silent 22 && npx supergateway --streamableHttp http://127.0.0.1:4242/mcp"
]
}
}
}
```
Things like `Take a snapshot of my browser.` should now work in your
Prompt Chat.
----
- SSE only for now, since we already have a http server with a port
there
- Upstream "page tests" can be executed over this CDP relay via
https://github.com/microsoft/playwright/pull/36286
- Limitations for now are everything what happens outside of the tab its
session is shared with -> `window.open` / `target=_blank`.
---------
Co-authored-by: Yury Semikhatsky <yurys@chromium.org>
Adds support for the new StreamableHttp transport. I'm not aware of any
clients that implement it, but somebody's gotta make the start! Once
some clients support it, we can also advertise it in the README.
