Parallel Browser Mcp
@etairl
关于 Parallel Browser Mcp
MCP server for parallel browser automation across multiple providers. Run multiple concurrent browser sessions in memory with a shared, provider-agnostic tool surface across Playwright (local Chromium), Browserbase, Anchor Browser, and Cloudflare Browser Run. Each session gets a
基本信息
配置
使用下面的配置,将此服务器添加到你的 MCP 客户端。
{
"mcpServers": {
"parallel-browser-mcp": {
"command": "npx",
"args": [
"parallel-browser-mcp@latest"
]
}
}
}工具
未检测到工具
工具是从 README 中自动提取的。维护者可以在 ## Tools 标题下列出工具,即可填充这部分内容。
概览
What is Parallel Browser Mcp?
Parallel Browser Mcp is an MCP server for parallel browser automation. It exposes a numeric session model so one client can create and control multiple browser sessions simultaneously across several providers: local Playwright Chromium, Browserbase, Anchor Browser, and Cloudflare Browser Run. It is designed for developers building AI-driven browser automation workflows within MCP‑compatible clients.
How to use Parallel Browser Mcp?
Install via npm (npx parallel-browser-mcp@latest) and add the server configuration to your MCP client (e.g., Claude, Cursor, VS Code). The server runs over stdio. Create sessions by calling start_session with a provider, then use any browser_* tool by passing the returned numeric sessionId. Provider credentials and defaults are set in environment variables or a JSON config object.
Key features of Parallel Browser Mcp
- Multiple concurrent browser sessions with numeric IDs in memory
- Unified provider abstraction for Playwright, Browserbase, Anchor, and Cloudflare
- Session management tools: start, close, close all, list sessions
- Over 20 browser automation tools (navigate, click, fill, screenshot, and more)
- Configuration via environment variables, config file, or inline JSON
Use cases of Parallel Browser Mcp
- Run multiple browser-based agents in parallel from a single MCP client
- Automate web scraping and form filling across different browser providers
- AI agents that need to manipulate several pages concurrently (e.g., research, data extraction)
- Test browser automation scripts locally with Playwright and switch to cloud providers
FAQ from Parallel Browser Mcp
Which providers does Parallel Browser Mcp support?
It supports four providers: Playwright (local Chromium), Browserbase (via @browserbasehq/sdk), Anchor (via anchorbrowser), and Cloudflare (via Cloudflare Browser Run).
How do I configure provider credentials?
Credentials are set via environment variables: BROWSERBASE_API_KEY, ANCHOR_API_KEY, CLOUDFLARE_API_TOKEN, and CLOUDFLARE_ACCOUNT_ID. Playwright requires no credentials. Additional provider‑specific options can be set in a JSON configuration provided through BROWSER_MCP_CONFIG or a config file.
How do I run the server?
You can run it locally with node dist/index.js after building, or use the npm CLI: npx parallel-browser-mcp@latest. The server communicates over stdio, so it integrates with any MCP client that supports stdio transport.
What tools are available for controlling a session?
The server offers session tools (start_session, close_session, close_all_sessions, get_sessions) and browser tools such as browser_navigate, browser_click, browser_fill, browser_screenshot, browser_snapshot, browser_evaluate, and more—over 20 in total.
How do I start and use a session?
Call start_session with a provider name (e.g., "playwright"). The response includes a numeric session ID. Then call any browser_* tool with { "sessionId": <id>, ...}. Finally, call close_session to release resources.
浏览器自动化 分类下的更多 MCP 服务器
browser-mcp
djydeA browser extension and MCP server that allows you to interact with the browser you are using.
Fastcrw
usFast, lightweight Firecrawl/Tavily alternative in Rust. Web scraper, crawler & search API with MCP server for AI agents. Drop-in Firecrawl-compatible API (/scrape, /crawl, /search). 2.3x faster than Tavily, 1.5x faster than Firecrawl in 1K-URL benchmarks. 6 MB RAM, single binary.
playwright-server MCP server
blackwhite084@mendableai/firecrawl Mcp Server
mendableaiThe API to search, scrape, and interact with the web at scale. 🔥
评论