browser-mcp
Allows AI agents to perform web browsing tasks using OpenAI's language models (e.g., GPT-4o-mini, o3-mini) through the browser-use library.
This server does not currently have a deployable release on Glama. Follow the repository's setup instructions to install and configure it for your MCP client, including any required credentials and local dependencies.
browser-mcp
A MCP (Model Control Protocol) server for browser-use library. This package allows AI agents to perform web browsing tasks through a standardized interface.
Installation
You can install the package using pip:
pip install browser-mcpOr with uv (recommended):
uv pip install browser-mcpAfter installation, you'll need to install Playwright's browser dependencies:
playwright installAlternatively, you can use the browser-mcp-run command which will automatically install these dependencies if they're missing.
Related MCP server: Cloudflare Playwright MCP
Setup
For development, clone the repository and install in development mode:
# Clone the repository
git clone https://github.com/pranav7/browser-mcp.git
cd browser-mcp
# Install dependencies with uv
uv pip install -e .
# Or with pip
pip install -e .Environment Variables
Create a .env file with your OpenAI API key:
OPENAI_API_KEY=your_api_key_hereUsage
Running the MCP Server
In Development Mode
When working with the package in development mode, you can run it directly with Python:
mcp dev browser_mcp/server.pyIn Production
After installing the package from PyPI, you can run it with uvx:
uvx browser-mcpThe package is specifically designed to work with uvx, which allows for more efficient package loading and execution.
With Automatic Dependency Check
You can also use the browser-mcp-run command, which checks for and installs Playwright dependencies automatically before starting the server:
browser-mcp-runThis ensures that all required Playwright browsers are installed on your system.
Using as a Client
from mcp.client import Client
async def main():
client = await Client.connect()
# Perform a task with the browser
result = await client.rpc("perform_task_with_browser",
task="Search for the latest news about AI and summarize the top 3 results")
print(result)
await client.close()Programmatic Usage
You can also use the package programmatically:
# In development mode
from src import run
# In production (after installing the package)
# from browser_mcp import run
# Run the MCP server with stdio transport
run(transport="stdio")
# Or with SSE transport
# run(transport="sse")Available RPC Methods
search_web(task: str, model: str = "gpt-4o-mini")- Performs basic web searches using browser-use Agent. Themodelparameter is optional and defaults to "gpt-4o-mini".search_web_with_planning(task: str, base_model: str = "gpt-4o-mini", planning_model: str = "o3-mini")- Performs complex web searches that require planning. Uses a planner LLM for better task decomposition. Bothbase_modelandplanning_modelparameters are optional with their respective defaults.
Development
Testing
Tests can be run with:
python -m unittest discoverYou can also test the package functionality with:
python test_uvx.pyThis script will:
Test importing the package directly (development mode)
Attempt to run it with uvx (production mode)
Note: The uvx test may fail in development mode unless the package is published to PyPI. This is expected behavior.
Publishing to PyPI
This project uses GitHub Actions to automatically publish to PyPI when a new release is created. The workflow:
Builds the package using uv
Publishes it to PyPI using trusted publishing
To create a new release:
Update the version in
pyproject.tomlCreate a new release on GitHub
The GitHub Action will automatically build and publish the package
License
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
This server cannot be deployed
Maintenance
Related MCP Connectors
Official Browserless MCP server: browser automation, scraping, screenshots and PDFs for AI agents.
Web scraping, browser automation, screenshots and AI data extraction for AI agents via MCP.
Control a real browser from AI agents: navigate, click, type, screenshot, read the page.
A paid remote MCP for AI agent browser approval MCP, built to return verdicts, receipts, usage logs,
Related MCP Servers
- FlicenseNot gradedqualityDmaintenanceA server that enables AI systems to browse, retrieve content from, and interact with web pages through the Model Context Protocol.1-
- AlicenseNot gradedqualityNot gradedmaintenanceA Model Control Protocol server that enables AI assistants to control a browser through tools for web automation tasks like navigation, typing, clicking, and taking screenshots.Apache 2.0
- AlicenseNot gradedqualityNot gradedmaintenanceA Model Control Protocol server that enables AI assistants to control a browser through tools for web automation tasks like navigation, typing, clicking, and taking screenshots.Apache 2.0
- AlicenseNot gradedqualityNot gradedmaintenanceA Model Control Protocol server that enables AI assistants to control a browser, allowing them to perform web automation tasks like navigation, typing, clicking, and taking screenshots.Apache 2.0