Docy MCP Server

Docy is a Model Context Protocol (MCP) server built to give AI assistants instant, reliable access to technical documentation. In modern software development, documentation is often scattered across multiple sites—React docs, Python libraries, internal wikis, and vendor APIs. Developers routinely face stale links, rate limits, or broken pages that slow down coding workflows and reduce the accuracy of AI-generated help. Docy eliminates these friction points by scraping configured documentation sites with crawl4ai and exposing the results through MCP tools, so an assistant can query up‑to‑date content without leaving its environment.

The server’s core value lies in real‑time, self‑hosted documentation retrieval. By hosting the service inside your own infrastructure, teams keep sensitive API references and internal docs secure while still benefiting from AI augmentation. Docy’s hot‑reload capability means new documentation sources can be added on the fly by editing a simple file—no restart required. Intelligent caching further reduces latency, ensuring that repeated queries hit the local cache while still refreshing stale pages behind the scenes. This blend of immediacy and freshness makes Docy an essential companion for AI assistants that need authoritative references to answer code questions, debug issues, or suggest best practices.

Key tools provided by Docy simplify common documentation tasks. enumerates all configured sources, allowing an assistant to discover what is available. pulls a full page’s content as markdown, giving the assistant direct access to code snippets and explanations. retrieves all hyperlinks from a page, enabling navigation through related topics or API endpoints. Corresponding MCP prompts—documentation_sources, documentation_page, and documentation_links—offer an alternative, more natural invocation style for developers who prefer prompt-based interaction.

Real‑world scenarios that benefit from Docy include: verifying library usage against the latest official docs, comparing implementation patterns across frameworks (e.g., React vs. Vue), and ensuring that internal coding standards align with external API contracts. In a CI/CD pipeline, an AI assistant could automatically fetch the latest documentation during code reviews, flagging deprecated calls or missing error handling. In IDEs like VS Code, Docy can power context‑aware tooltips that surface up‑to‑date docs directly inside the editor, reducing context switching for developers.

Because Docy is built on FastMCP v2, it integrates seamlessly with any MCP‑enabled AI tool—Claude, VS Code extensions, or custom chatbots. By adding simple guidelines to a project’s , teams can enforce that documentation queries route through Docy instead of generic web fetch tools, ensuring consistent quality and compliance. This tight integration, combined with self‑hosted control, hot‑reload support, and intelligent caching, gives Docy a distinctive edge over ad‑hoc web scraping or third‑party documentation APIs.