Plasmate MCP Server

@plasmate-labs

代码仓库主页

基本信息

分类

开发工具

许可证

Apache-2.0

传输方式

stdio

作者

plasmate-labs

提交者

DBH

关于 Plasmate MCP Server

MCP server for Plasmate. Use with Claude Code, Claude Desktop, Cursor, and Windsurf. npm install plasmate-mcp.

#plasmate#browser#headless#free #opensource#apache#chrome #cdp#som

Plasmate MCP Server 配置

使用下面的配置,将此服务器添加到你的 MCP 客户端。

{
  "mcpServers": {
    "plasmate": {
      "command": "plasmate",
      "args": [
        "mcp"
      ]
    }
  }
}

Plasmate MCP Server 工具

Fetch a URL and get the Semantic Object Model (SOM)

Get clean readable text from any URL

Open a persistent browser session

Click elements by SOM ID

Execute JavaScript in the page

Close a browser session

概览

What is Plasmate Mcp Server?

Plasmate Mcp Server is an MCP server for Plasmate, a browser engine designed for AI agents. It enables Claude Code, Claude Desktop, Cursor, Windsurf, and any MCP client to browse the web through Plasmate instead of Chrome.

How to use Plasmate Mcp Server?

Install the wrapper globally with npm install -g plasmate-mcp and the Plasmate engine via pip install plasmate, cargo install plasmate, or brew install plasmate. Then add configuration to your MCP client (e.g., Claude Code settings or Claude Desktop config) using npx plasmate-mcp as the command.

Key features of Plasmate Mcp Server

Semantic Object Model (SOM) reduces tokens by 16.6×.
50× faster than Chrome with no rendering overhead.
Memory usage of only 30MB vs Chrome’s 300MB+.
Provides structured JSON output, not raw HTML.
Tools: fetch_page, extract_text, open_page, click, evaluate, close_page.
Works with multiple MCP clients: Claude Code, Claude Desktop, Cursor, Windsurf.

Use cases of Plasmate Mcp Server

AI agents browse the web with far lower token consumption.
Extract clean readable text from any URL for LLM ingestion.
Interact with web pages programmatically (click, evaluate JS).
Run persistent browser sessions for multi-step web tasks.
Replace Chrome for agentic browsing to reduce cost and latency.

常见问题

How does Plasmate compare to using Chrome?

Plasmate uses a Semantic Object Model (SOM) that is 16.6× fewer tokens, 50× faster, and uses only 30MB of memory compared to Chrome’s 300MB+.

What are the runtime dependencies?

You must install both the `plasmate-mcp` npm package and the Plasmate engine (via pip, cargo, or Homebrew).

Where does data live?

Browsing sessions and fetched page data are managed by the Plasmate engine; state is not persisted across server restarts unless you manually manage sessions.

What transport does the server use?

It uses standard MCP stdio transport configured via command-line invocation (e.g., `npx plasmate-mcp`).

开发工具分类下的更多 MCP 服务器

OrangePro

Aamir Siddiqui

OrangePro local-first CLI + MCP server for behavior mapping, grounded test generation, and dynamic proof.

Eddyter AI Rich Text Editor MCP

EddyterAI

Set up and manage the Eddyter AI rich text editor by talking to your AI agent — for Claude, Cursor, Codex and other MCP tools. A Model Context Protocol server that lets AI tools add Eddyter to a React, Next.js, Vite or R

Yocoolab

MCP server for AI Agents — exposes Yocoolab feedback threads and design selections as tools

Capgo MCP

Cap-go

Console, Backend and CLI to manage Capgo Instant update and Native build for Capacitor apps Build and upload to app store your mobile app