What is the default Ollama model used?
The default model is `qwen3`. You can change it by setting the `OLLAMA_MODEL` environment variable or passing the `--model` flag to the client.
@Cam10001110101
基本信息
暂无概览
使用下面的配置,将此服务器添加到你的 MCP 客户端。
{
"mcpServers": {
"mcp-server-browser-use-ollama": {
"command": "uv",
"args": [
"pip",
"install",
"-e",
"."
]
}
}
}工具列表暂未提供
当前收录暂时没有可展示的工具列表,不代表服务没有工具。请查阅项目文档了解支持的能力。
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。
Browser Use MCP is a browser automation system that enables AI agents to control web browsers through the Model Context Protocol (MCP). It is designed to work with local AI models running via Ollama, allowing secure and efficient browser interactions using locally-hosted LLMs.
Clone the repository, install dependencies with uv pip install -e . and run playwright install. Ensure Ollama is installed, running (ollama serve), and has a model pulled (e.g., ollama pull qwen3). Use the server via Claude Desktop by configuring the claude_desktop_config.json to point to src/server.py, or run the client directly with python src/client.py src/server.py [task].
The default model is `qwen3`. You can change it by setting the `OLLAMA_MODEL` environment variable or passing the `--model` flag to the client.
It supports Chrome, Firefox, and Safari through Playwright automation.
Browser sessions are automatically cleaned up, and the system includes graceful recovery from parsing errors, with comprehensive logging for debugging.
It uses a client‑server architecture with stdio‑based MCP transport: the user interacts with a client, which communicates via MCP with the server, which in turn controls the browser via Playwright.
Yes, any Ollama‑compatible model can be used. The system is optimized for coding models like `qwen3` and `qwen2.5-coder:7b`, and model selection is configurable via environment variables or command‑line arguments.
Allow LLMs to control a browser with Browserbase and Stagehand
An MCP server for fetching URLs / Youtube video transcript.
Fast, 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.

Model Context Protocol Servers

MCP server for managing Gologin browser profiles, proxies, fingerprints and cloud browser sessions through AI conversations
评论