Steel MCP Server
by steel-dev
The Steel MCP Server enables LLMs like Claude to navigate the web using Puppeteer-based tools and Steel. It provides tools for standard web actions like clicking, scrolling, typing, and taking screenshots, based on the Web Voyager framework.
Last updated: N/A
What is Steel MCP Server?
The Steel MCP Server is a Model Context Protocol server that allows Large Language Models (LLMs) like Claude to interact with and navigate the web. It leverages Puppeteer for browser automation and Steel for session management, providing a suite of tools for performing various web-based tasks.
How to use Steel MCP Server?
To use the Steel MCP Server, you need to configure Claude Desktop with the server details, including the command to run the server and environment variables for either local or cloud mode. You can then interact with the server through Claude, asking it to perform tasks like searching, navigating, clicking, and filling out forms. Detailed instructions are provided in the Quick Start section of the README.
Key features of Steel MCP Server
Browser automation with Puppeteer
Steel integration for browser session management
Visual element identification through numbered labels
Screenshot capabilities
Basic web interaction (navigation, clicking, form filling)
Lazy-loading support through scrolling
Local and remote Steel instance support
Use cases of Steel MCP Server
Search for a recipe and save the ingredients list
Track a package delivery status
Find and compare prices for a specific product
Fill out an online job application
FAQ from Steel MCP Server
How do I run the server in cloud mode?
How do I run the server in cloud mode?
Set STEEL_LOCAL to "false" and provide a valid STEEL_API_KEY in the environment variables.
How do I run the server locally?
How do I run the server locally?
Set STEEL_LOCAL to "true" and ensure your local Steel service is running. Optionally, set STEEL_BASE_URL to point to your local Steel server.
What is the purpose of the numbered labels on the web page?
What is the purpose of the numbered labels on the web page?
The numbered labels identify interactive elements (buttons, links, inputs) for click and type operations.
How do I capture a screenshot?
How do I capture a screenshot?
Use the 'save_unmarked_screenshot' tool and specify a resource name. The screenshot will be stored as a resource accessible via an MCP resource URI.
Why is Claude slowing down after several browser actions?
Why is Claude slowing down after several browser actions?
After ~15-20 browser actions Claude starts to slow down as it's context window gets filled with images. There may be some latency, especially with the Claude Desktop client lagging behind.