Steel Voyager FAQs

Question 1

What is Steel Voyager?

Accepted Answer

Steel Voyager is a Model Context Protocol (MCP) server that lets LLMs like Claude control and automate web browsers using Puppeteer and the Steel platform. It enables AI agents to perform tasks such as web searching, data extraction, and form filling.

Question 2

How does Steel Voyager work with Claude?

Accepted Answer

Steel Voyager acts as a bridge between Claude and a web browser. It exposes web browsing actions as tools that Claude can use, such as navigating to URLs, clicking elements, typing text, and taking screenshots, all controlled via MCP.

Question 3

What are the key features of Steel Voyager?

Accepted Answer

Key features include browser automation with Puppeteer, Steel integration for session management, visual element identification through numbered labels, screenshot capabilities, support for both local and remote Steel instances, and tools for basic web interactions.

Question 4

Do I need a Steel API key to use Steel Voyager?

Accepted Answer

Yes, you'll need a Steel API key if you are running in cloud mode (STEEL_LOCAL="false"). If you're running a local or self-hosted Steel instance (STEEL_LOCAL="true"), an API key is not required.

Question 5

How do I identify elements on a webpage for clicking or typing?

Accepted Answer

Steel Voyager adds visual overlays to webpages with numbered labels on interactive elements. These numbers correspond to elements that Claude can then interact with via the 'click' or 'type' tools.

Steel Voyager

关于

主要功能

使用案例