docs-mcp

loading models…
search
hybrid blends meaning and exact words · vector favors meaning · keyword favors exact terms · / jumps back here
filters
try a query above, or narrow it with mode & filters first
indexed sources
loading…
add documentation
only enable this if depth and max pages cover the whole site — otherwise pages outside the crawl's reach get deleted

drop files here or click to browse

HTML, Markdown, PDF, TXT
local server only — the path must be accessible from the machine running docs-mcp

drop requirements.txt, pyproject.toml, or package.json

auto-discovers documentation URLs for each dependency
system
Embedding Model —
Embedding Dims —
LLM Model —
Python Version —
Indexed Sources —
Total Pages —
Total Chunks —
about

Documentation RAG system: crawl a docs site, upload files, or index a local folder — then semantic search and chat with your docs via MCP, REST API, or web UI.

github · pypi · mcp protocol