Plasmate MCP
Agent-native headless browser that converts web pages to structured Semantic Object Model (SOM) JSON -- 4x fewer tokens than raw HTML with lower latency on Claude and GPT-4o.
plasmate-mcp
MCP server for Plasmate - the browser engine for AI agents.
Gives Claude Code, Claude Desktop, Cursor, Windsurf, and any MCP client the ability to browse the web through Plasmate instead of Chrome.
Install
# Install the MCP wrapper
npm install -g plasmate-mcp
# Install the Plasmate engine (pick one)
pip install plasmate
# or: cargo install plasmate
# or: brew tap plasmate-labs/plasmate && brew install plasmate
Use with Claude Code
claude mcp add plasmate -- npx plasmate-mcp
Or manually add to ~/.claude/settings.json:
{
"mcpServers": {
"plasmate": {
"command": "npx",
"args": ["-y", "plasmate-mcp"]
}
}
}
Use with Claude Desktop
Add to claude_desktop_config.json:
{
"mcpServers": {
"plasmate": {
"command": "npx",
"args": ["-y", "plasmate-mcp"]
}
}
}
Use with Cursor
Add to Cursor's MCP settings:
{
"mcpServers": {
"plasmate": {
"command": "npx",
"args": ["-y", "plasmate-mcp"]
}
}
}
Available Tools
Once connected, your AI gets these tools:
| Tool | Description |
|---|---|
fetch_page | Fetch a URL and get the Semantic Object Model (SOM) |
extract_text | Get clean readable text from any URL |
open_page | Open a persistent browser session |
click | Click elements by SOM ID |
evaluate | Execute JavaScript in the page |
close_page | Close a browser session |
Why Plasmate instead of Chrome?
- 16.6x fewer tokens - SOM is dramatically smaller than DOM dumps
- 50x faster - No rendering engine overhead
- 30MB memory - vs Chrome's 300MB+
- Structured output - Clean JSON, not raw HTML
Links
Part of the Plasmate Ecosystem
| Engine | plasmate - The browser engine for agents |
| MCP | plasmate-mcp - Claude Code, Cursor, Windsurf |
| Extension | plasmate-extension - Chrome cookie export |
| SDKs | Python / Node.js / Go / Rust |
| Frameworks | LangChain / CrewAI / AutoGen / Smolagents |
| Tools | Scrapy / Audit / A11y / GitHub Action |
| Resources | Awesome Plasmate / Notebooks / Benchmarks |
| Docs | docs.plasmate.app |
| W3C | Web Content Browser for AI Agents |
Máy chủ liên quan
Bright Data
nhà tài trợDiscover, extract, and interact with the web - one interface powering automated access across the public internet.
yt-dlp
Download video and audio from YouTube and other platforms using the yt-dlp tool.
DOMShell
Browse the web with filesystem commands. 38 MCP tools let AI agents ls, cd, grep, click, and type through Chrome via a Chrome Extension.
Outscraper
Extract data from Google Maps, including places and reviews, using the Outscraper API.
Website to Markdown MCP Server
Fetches and converts website content to Markdown with AI-powered cleanup, OpenAPI support, and stealth browsing.
Cloudflare Browser Rendering
Provides web context to LLMs using Cloudflare's Browser Rendering API.
Playwright MCP
Automate web interactions and perform web scraping tasks using the Playwright framework.
Web Search
Performs web searches and extracts full page content from search results.
MCP YouTube Transcript Server
Retrieves transcripts from YouTube videos for content analysis and processing.
Tech Collector MCP
Collects and summarizes technical articles from sources like Qiita, Dev.to, NewsAPI, and Hacker News using the Gemini API.
Secure Fetch
Secure fetch to prevent access to local resources