# browser-use: AI Browser Agents and Stealth Browsers for MCP Clients

> browser-use is a Python library and cloud service for AI browser agents, offering stealth browsers and a hosted agent API. It solves the problem of programmatic web interaction for AI, making it a key MCP Client for developers building intelligent automation.

**Published:** 2026-10-20T12:01:52.771+00:00

**Keywords:** browser-use,mcp-client,ai-browser-agents

# browser-use: Driving AI Browser Agents with MCP

browser-use provides an open-source Python library designed for AI browser agents, complemented by a cloud service offering stealth browsers and a hosted agent API. This makes it a direct fit for developers leveraging the Model Context Protocol (MCP) to integrate advanced AI capabilities into their projects, particularly those requiring sophisticated web interaction.

## Core Capabilities: AI Browser Agents and Stealth Browsing

The primary offering of browser-use centers around its AI browser agents. These agents are built to navigate the web in a manner that mirrors human interaction, allowing for complex, dynamic web operations. For developers, this means the ability to programmatically control browser actions through an AI, opening up possibilities for data extraction, automated testing, or interactive task completion on the web.

A significant feature supporting these agents is the provision of "stealth browsers." These are essential for maintaining operational integrity in scenarios where standard automated browsing might encounter detection or blocking mechanisms. The combination of AI-driven control and stealth capabilities ensures that agents can operate effectively across a wide range of web environments.

## Cloud-Hosted Agent API

Beyond the local Python library, browser-use offers a "Browser Use Cloud" service. This cloud component includes a hosted agent API, which streamlines the deployment and management of AI browser agents. Instead of self-hosting and maintaining browser environments, developers can leverage this API to run their agents remotely. This setup is particularly beneficial for scaling operations or for teams that prefer a managed service approach to their AI browser agent infrastructure. The hosted API simplifies the integration process, allowing MCP Clients to dispatch tasks and receive results without deep browser-level management.

## References
- [browser-use on GitHub](https://github.com/browser-use/browser-use)
- [Model Context Protocol Documentation](https://modelcontextprotocol.io/introduction)
- [browser-use on model-context-protocol.com](https://model-context-protocol.com/clients/)

## Related Repository

- [browser-use](https://model-context-protocol.com/clients/browser-use)

**Source:** https://model-context-protocol.com/blog/browser-use-ai-browser-agents-and-stealth-browsers-for-mcp-mcp-client-guide