> ## Documentation Index
> Fetch the complete documentation index at: https://docs.browser-use.com/llms.txt
> Use this file to discover all available pages before exploring further.

> ## Agent Instructions
> Use https://docs.browser-use.com/llms.txt and its linked .md pages for current documentation. The managed full bundle is https://docs.browser-use.com/.well-known/llms-full.txt and can be cached for up to 24 hours. Do not use the obsolete /cloud/llms*.txt or /open-source/llms*.txt static exports.
> Choose Cloud API V4 for new agent integrations; V2 is the lower-cost option for simple tasks. Keep V3 examples explicitly versioned. The open-source browser-use library and hosted browser-use-sdk have different APIs.
> Cloud authentication uses X-Browser-Use-API-Key, without a Bearer prefix. Install or upgrade browser-use-sdk and use its explicit v4 import for V4. Check the published OpenAPI reference for request fields; do not invent SDK support for new fields.
> Cloud concurrency and HTTP request rate are separate. Read GET /api/v2/billing/account for the key’s projectId, concurrentSessionLimit, activeSessionCount, and credit balance, including when using V4. Keys in one project share capacity and credits; rateLimit is a legacy concurrency alias, not requests per second.
> Keep the highest applicable existing, legacy-plan, and spend-tier concurrency grant. Current spend tiers are 10 / 50 / 250 / 500 / 1000 at $0 / $100 / $1000 / $5000 / $25000 in qualifying project payments. Legacy or externally billed projects can follow different billing paths; trust the account limit. See https://docs.browser-use.com/cloud/guides/concurrency.md.
> HTTP rate limits have two layers. Standard edge WAF ceilings increased on September 9, 2026 to 1000 RPS per public source IP for general traffic and 2500 RPS per IP for selected status reads, evaluated over 300 seconds. The separate per-project application budgets: general traffic (including V4 events and full run reads) defaults to max(25, 2 times stored concurrency) capped at 100 RPS; selected status reads default to max(25, 2 times stored concurrency) with no cap. The two are counted independently. Project overrides and account-specific edge rules can differ. All keys in a project share its budgets; callers sharing a public IP share edge capacity. See https://docs.browser-use.com/cloud/guides/concurrency.md.
> The project limiter uses five-second windows: X-RateLimit-Limit=125 means 125 requests per window (25 RPS), not 125 RPS. Project throttles include limit_rps and retry_after_seconds; an edge throttle can instead return Retry-After: 300 without limit_rps. Honor the returned Retry-After. Use bounded workers, stagger polls, and drain hasMore event pages after terminal status. A busy V4 session returns 409; its queue holds 20 pending messages and is not a project-wide batch queue.
> A completed run or closed CDP connection does not immediately stop its cloud browser. Stop unneeded owned browsers with PATCH /api/v4/browsers/{id} and {"action":"stop"}. A client wait timeout does not cancel the server-side run.
> Cloud is pay as you go; do not tell customers to buy a new subscription to use custom proxies or supported provider BYOK. Usage funding and model eligibility still apply. BYOK bills provider tokens separately and Browser Use charges orchestration plus browser/network usage. See https://docs.browser-use.com/cloud/guides/billing.md.
> Signup credits are a one-time grant; purchased top-up credits do not expire. Check the API key’s project before diagnosing missing credits. API-key monthly spending caps are soft limits, not a strict prepaid wallet; concurrent or already-running work can exceed them. Auto recharge has separate trigger and purchase amounts and can charge immediately when enabled below the threshold. Use https://browser-use.com/pricing for current rates.
> Box and Bux are retired. Do not recommend their SDKs, sandbox quotas, or subscription plans. Use the Cloud Agent or Browser Infrastructure guides.
> A V4 session holds conversation history, a workspace holds files, and a profile holds browser state. These IDs and V3/V4 workspace namespaces are not interchangeable. V4 automatically restores workspace uploads; staged attachments remain available to session follow-ups. Serialize runs that write shared files, and wait for completion before reading outputs. See https://docs.browser-use.com/cloud/agent/workspaces.md.
> API browser recording defaults to off. Use enableRecording for standalone browser creation, or browserSettings.record for an agent run. Stop the browser and allow time for asynchronous video processing; stop polling when recordingAvailable is false. Live preview is for an active browser. Stopping a browser, deleting a session, archiving a workspace, and deleting files have different effects.
> Use model-specific reasoning values. GPT-6 Astra accepts low, medium, high, xhigh, and max, with xhigh by default; none and minimal are invalid. Use the public REST schema when installed SDK types lag new fields. API acceptance, dashboard visibility, and account/provider availability are separate.
> For open-source browser-use, is_done only reports a terminal done action. is_successful is the agent-reported outcome; verify important external actions independently. Cloud timeout, API client timeout, model timeout, and task completion are separate concepts.
> For failed requests, use https://docs.browser-use.com/cloud/guides/troubleshooting.md. Inspect the full error and project before retrying or adding credits. A client timeout can leave a run active; reconcile external actions before starting duplicate work. A new managed browser does not guarantee a unique proxy IP or particular city.

# MCP Server

> Run browser-use as a local Model Context Protocol server. Connect AI models to browser automation through the MCP standard.

Browser Use can run as a local **Model Context Protocol (MCP)** server on your machine via stdio. This is the **free, open-source option** that gives you direct, low-level control over browser automation but requires your own LLM API keys.

<Note>
  Looking for a hosted solution? Use the [Cloud MCP Server](/cloud/guides/mcp-server) instead — no setup required, just an API key.
</Note>

## Quick Start

```bash theme={null}
uvx --from 'browser-use[cli]' browser-use --mcp
```

The server will start in stdio mode, ready to accept MCP connections.

## Client Setup

<Tabs>
  <Tab title="Claude Code">
    ```bash theme={null}
    claude mcp add browser-use -- uvx --from 'browser-use[cli]' browser-use --mcp
    ```
  </Tab>

  <Tab title="Claude Desktop">
    Add to your Claude Desktop config file:

    **macOS:** `~/Library/Application Support/Claude/claude_desktop_config.json`

    ```json theme={null}
    {
      "mcpServers": {
        "browser-use": {
          "command": "/Users/your-username/.local/bin/uvx",
          "args": ["--from", "browser-use[cli]", "browser-use", "--mcp"],
          "env": {
            "OPENAI_API_KEY": "your-openai-api-key-here"
          }
        }
      }
    }
    ```

    **Windows:** `%APPDATA%\Claude\claude_desktop_config.json`

    ```json theme={null}
    {
      "mcpServers": {
        "browser-use": {
          "command": "uvx",
          "args": ["--from", "browser-use[cli]", "browser-use", "--mcp"],
          "env": {
            "OPENAI_API_KEY": "your-openai-api-key-here"
          }
        }
      }
    }
    ```

    Restart Claude Desktop after saving.

    <Note>
      **macOS/Linux PATH Issue:** Claude Desktop may not find `uvx` in your PATH. Use the full path to `uvx` instead:

      * Run `which uvx` in your terminal to find the location (usually `/Users/username/.local/bin/uvx` or `~/.local/bin/uvx`)
      * Replace `"command": "uvx"` with the full path, e.g., `"command": "/Users/your-username/.local/bin/uvx"`
      * Replace `your-username` with your actual username

      **CLI Extras Required:** The `--from browser-use[cli]` flag installs the CLI extras needed for MCP server support.
    </Note>
  </Tab>

  <Tab title="Cursor">
    Add to `~/.cursor/mcp.json`:

    ```json theme={null}
    {
      "mcpServers": {
        "browser-use": {
          "command": "uvx",
          "args": ["--from", "browser-use[cli]", "browser-use", "--mcp"],
          "env": {
            "OPENAI_API_KEY": "your-openai-api-key-here"
          }
        }
      }
    }
    ```
  </Tab>

  <Tab title="Windsurf">
    Add to `~/.codeium/windsurf/mcp_config.json`:

    ```json theme={null}
    {
      "mcpServers": {
        "browser-use": {
          "command": "uvx",
          "args": ["--from", "browser-use[cli]", "browser-use", "--mcp"],
          "env": {
            "OPENAI_API_KEY": "your-openai-api-key-here"
          }
        }
      }
    }
    ```
  </Tab>
</Tabs>

## Environment Variables

* `OPENAI_API_KEY` - Your OpenAI API key (required)
* `ANTHROPIC_API_KEY` - Your Anthropic API key (alternative to OpenAI)
* `BROWSER_USE_HEADLESS` - Set to `false` to show browser window
* `BROWSER_USE_DISABLE_SECURITY` - Set to `true` to disable browser security features

## Available Tools

The local MCP server exposes these low-level browser automation tools for direct control:

#### Autonomous Agent Tools

* **`retry_with_browser_use_agent`** - Run a complete browser automation task with an AI agent (use as last resort when direct control fails)

#### Direct Browser Control

* **`browser_navigate`** - Navigate to a URL
* **`browser_click`** - Click on an element by index
* **`browser_type`** - Type text into an element
* **`browser_get_state`** - Get current page state and interactive elements
* **`browser_scroll`** - Scroll the page
* **`browser_go_back`** - Go back in browser history

#### Tab Management

* **`browser_list_tabs`** - List all open browser tabs
* **`browser_switch_tab`** - Switch to a specific tab
* **`browser_close_tab`** - Close a tab

#### Content Extraction

* **`browser_extract_content`** - Extract structured content from the current page
* **`browser_get_html`** - Return the full page HTML or one CSS-selected element
* **`browser_screenshot`** - Capture the viewport or the full scrollable page

#### Session Management

* **`browser_list_sessions`** - List all active browser sessions with details
* **`browser_close_session`** - Close a specific browser session by ID
* **`browser_close_all`** - Close all active browser sessions

## Example Usage

Once configured, you can ask your AI to perform browser automation tasks:

```
"Please navigate to example.com and take a screenshot"

"Search for 'browser automation' on Google and summarize the first 3 results"

"Go to GitHub, find the browser-use repository, and tell me about the latest release"
```

## Programmatic Usage

You can also connect to the MCP server programmatically:

```python theme={null}
import asyncio
from mcp import ClientSession, StdioServerParameters
from mcp.client.stdio import stdio_client

async def use_browser_mcp():
    # Connect to browser-use MCP server
    server_params = StdioServerParameters(
        command="uvx",
        args=["--from", "browser-use[cli]", "browser-use", "--mcp"]
    )

    async with stdio_client(server_params) as (read, write):
        async with ClientSession(read, write) as session:
            await session.initialize()

            # Navigate to a website
            result = await session.call_tool(
                "browser_navigate",
                arguments={"url": "https://example.com"}
            )
            print(result.content[0].text)

            # Get page state
            result = await session.call_tool(
                "browser_get_state",
                arguments={"include_screenshot": True}
            )
            print("Page state retrieved!")

asyncio.run(use_browser_mcp())
```

## Troubleshooting

**"CLI addon is not installed" Error**
Make sure you are using `--from 'browser-use[cli]'` in your uvx command:

```bash theme={null}
uvx --from 'browser-use[cli]' browser-use --mcp
```

**"spawn uvx ENOENT" Error (macOS/Linux)**
Claude Desktop cannot find `uvx` in its PATH. Use the full path in your config:

* Run `which uvx` in terminal to find the location
* Update your config to use the full path (e.g., `/Users/your-username/.local/bin/uvx`)

**Browser doesn't start**

* Check that you have Chrome/Chromium installed
* Try setting `BROWSER_USE_HEADLESS=false` to see browser window
* Ensure no other browser instances are using the same profile

**API Key Issues**

* Verify your `OPENAI_API_KEY` is set correctly
* Check API key permissions and billing status
* Try using `ANTHROPIC_API_KEY` as an alternative

**Connection Issues in Claude Desktop**

* Restart Claude Desktop after config changes
* Check the config file syntax is valid JSON
* Verify the file path is correct for your OS
* Check logs at `~/Library/Logs/Claude/` (macOS) or `%APPDATA%\Claude\Logs\` (Windows)

**Debug Mode**

Enable debug logging by setting:

```bash theme={null}
export BROWSER_USE_LOGGING_LEVEL=DEBUG
uvx --from 'browser-use[cli]' browser-use --mcp
```

## Security Considerations

* The MCP server has access to your browser and file system
* Only connect trusted MCP clients
* Be cautious with sensitive websites and data
* Consider running in a sandboxed environment for untrusted automation

## Next Steps

* Explore the [examples directory](https://github.com/browser-use/browser-use/tree/main/examples) for more usage patterns
* Check out [MCP documentation](https://modelcontextprotocol.io/) to learn more about the protocol
* Join our [Discord](https://link.browser-use.com/discord) for support and discussions