Frequently Asked Questions (FAQ) about Docy

Q: What is Docy? A: Docy is a Model Context Protocol (MCP) server that provides AI agents with real-time access to technical documentation. It allows AI models to search, retrieve, and navigate documentation from various sources.

Q: How does Docy integrate with UBOS? A: Docy seamlessly integrates with the UBOS AI Agent Development Platform, enhancing the capabilities of AI agents by providing them with accurate and up-to-date documentation access.

Q: What are the key benefits of using Docy? A: Key benefits include instant documentation access, hot-reload support, intelligent caching, self-hosted control, and seamless MCP integration.

Q: What type of documentation can Docy access? A: Docy can access documentation from popular tech stacks like React, Python, and crawl4ai, as well as any other technology for which documentation is available online.

Q: How do I add new documentation sources to Docy? A: You can add new documentation sources by editing the .docy.urls file and adding the URLs of the documentation websites you want to include. The server supports hot-reloading, so no restart is needed.

Q: What is MCP? A: MCP stands for Model Context Protocol. It is an open protocol that standardizes how applications provide context to LLMs, enabling AI models to access and interact with external data sources and tools.

Q: What is crawl4ai? A: crawl4ai is a web scraping tool used by Docy to crawl and index documentation websites, ensuring that the documentation content is up-to-date.

Q: How does Docy’s caching mechanism work? A: Docy uses intelligent caching to reduce latency and minimize external requests. It pre-fetches and caches documentation URLs at startup, and the cache TTL can be configured via the DOCY_CACHE_TTL environment variable.

Q: How can I configure Docy? A: Docy can be configured using environment variables or a .docy.urls file. Environment variables allow you to customize various settings, such as the documentation URLs, cache TTL, and cache directory.

Q: What is the recommended way to install Docy? A: The recommended way to install Docy is using uv. Alternatively, you can use pip or Docker.

Q: How do I ensure that Claude Code prioritizes Docy for documentation-related tasks? A: Add specific guidelines to your project’s CLAUDE.md file, instructing Claude Code to prefer using Docy over WebFetchTool for documentation tasks.

Q: What are the available tools in Docy? A: Available tools include list_documentation_sources_tool, fetch_documentation_page, and fetch_document_links.

Q: How do I use the MCP inspector to debug Docy? A: You can use the MCP inspector with Docy using the command DOCY_DOCUMENTATION_URLS="https://docs.crawl4ai.com/" npx @modelcontextprotocol/inspector uvx mcp-server-docy.

Q: Is Docy open source? A: Yes, mcp-server-docy is licensed under the MIT License.