Frequently Asked Questions about Steel MCP Server

Q: What is the Steel MCP Server?

A: The Steel MCP Server is a Model Context Protocol (MCP) server that enables Large Language Models (LLMs) like Claude to interact with the web through Puppeteer-based tools and Steel. It allows LLMs to navigate, search, interact with, and extract information from websites.

Q: How does the Steel MCP Server work?

A: It acts as a bridge between LLMs and the web. It translates instructions from LLMs into actionable web commands, leveraging Puppeteer to automate browser interactions.

Q: What are the main use cases for the Steel MCP Server?

A: Some use cases include automated research, price comparison, data extraction, form filling, package tracking, and recipe retrieval.

Q: What are the key features of the Steel MCP Server?

A: Key features include browser automation with Puppeteer, Steel integration for session management, visual element identification, screenshot capabilities, basic web interaction support, lazy-loading support, and local/remote Steel instance support.

Q: What is MCP (Model Context Protocol)?

A: MCP is an open protocol that standardizes how applications provide context to LLMs, allowing AI models to access and interact with external data sources and tools.

Q: What is Puppeteer?

A: Puppeteer is a Node library that provides a high-level API to control Chrome or Chromium over the DevTools Protocol. It can be used to automate browser tasks, generate screenshots, and more.

Q: How do I configure the Steel MCP Server?

A: The server can be configured to run in either local or cloud mode using environment variables like STEEL_LOCAL, STEEL_API_KEY, and STEEL_BASE_URL.

Q: What is the difference between local and cloud modes?

A: In local mode, the server connects to a locally running Steel instance. In cloud mode, it connects to the Steel cloud service.

Q: How do I install the Steel MCP Server?

A: It can be installed via Smithery or by cloning the repository from GitHub and building it locally.

Q: How can I use the Steel MCP Server with Claude Desktop?

A: You need to configure the Claude Desktop configuration file (~/Library/Application Support/Claude/claude_desktop_config.json) with the appropriate command and environment variables to launch the server.

Q: What is UBOS and how does it relate to the Steel MCP Server?

A: UBOS is a full-stack AI Agent Development Platform. The Steel MCP Server integrates with UBOS, enabling AI Agents to perform web-based tasks.

Q: What are the benefits of using the Steel MCP Server with UBOS?

A: It enables AI Agents to perform complex research tasks, automate business processes, and provide personalized customer service through web interfaces.

Q: What if I encounter issues while using the Steel MCP Server?

A: Common troubleshooting steps include verifying your Steel API key, ensuring your local Steel instance is running, checking network connectivity, and adjusting viewport size settings.

Q: How can I contribute to the Steel MCP Server project?

A: You can contribute by forking the repository, creating a feature branch, and submitting a pull request.

Q: Is the Steel MCP Server production-ready?

A: The project is experimental and under active development. Use in production environments at your own risk.

Q: What are bounding boxes and how are they used?

A: When interacting with pages, Steel Puppeteer adds visual overlays to help identify interactive elements such as buttons, links, inputs using a unique numbered label.