Docs Converter FAQs

Question 1

What is Docs Converter?

Accepted Answer

Docs Converter transforms any documentation URL into a ready-to-run Model Context Protocol (MCP) server. This enables AI clients like Claude Desktop or Cursor to perform semantic search, source retrieval, and more on your documentation.

Question 2

Does Docs Converter require API keys or external services?

Accepted Answer

No, by default. Docs Converter operates 100% locally using Transformers.js for embeddings, requiring no API keys. You only need Node.js, Docker for ChromaDB, and Playwright browsers. OpenAI is an optional, opt-in embedding provider.

Question 3

How does Docs Converter process documentation?

Accepted Answer

It follows a comprehensive pipeline: It crawls a specified URL, cleans HTML, converts content to Markdown, chunks it, generates embeddings, stores them in a vector store (ChromaDB), and then exposes these through an MCP server.

Question 4

What tools does the MCP server expose?

Accepted Answer

The MCP server provides three key tools: `search_docs(query, topK?)` for semantic search, `get_source(url)` to retrieve all chunks from a specific documentation URL, and `list_sources()` to list all indexed documentation sources.

Question 5

Can I use Docs Converter with my preferred AI agent or LLM IDE?

Accepted Answer

Yes, Docs Converter generates an MCP server that is compatible with MCP clients. Configuration examples are provided for popular tools like Claude Desktop and Cursor, allowing them to connect and utilize your documentation server.

Docs Converter

Docs Converter

주요 기능

사용 사례

주요 기능

사용 사례