refdocs
Enables searching GitHub alongside public technical documentation to find facts or code snippets behind documentation questions, returning matching results with titles, URLs, and content or snippets when available.
This server does not currently have a deployable release on Glama. Follow the repository's setup instructions to install and configure it for your MCP client, including any required credentials and local dependencies.
refdocs
refdocs answers documentation questions and turns web pages into something an agent can actually read. Search runs across public technical documentation and GitHub for the facts or code snippet behind a question, and any URL can be fetched and handed back as clean markdown instead of raw HTML.
It is one small program with two faces: a command-line tool for quick lookups, and a stdio MCP server that gives agents the same two operations as tools.
Configuration
Both operations require a Ref API key:
export REF_API_KEY="your-key-here"The search endpoint defaults to https://api.ref.tools. Override it
with REF_API_BASE_URL when you are pointing at another deployment.
The key is checked before any request is made; when it is missing the
tool fails with an error that names REF_API_KEY.
Related MCP server: Web Research MCP
Install
npx @reneturcios/refdocs search "How do I stream a response with fetch?"Or run the container image:
docker run -i -e REF_API_KEY ghcr.io/cloudcorprecords/refdocsThe container defaults to the MCP server (see below). Pass CLI arguments to use it as a CLI instead.
CLI
# Search the documentation
refdocs search "How do I use the cache API in Next.js?"
refdocs search "What does this error mean?" --json
# Read a page as markdown
refdocs read https://example.test/some/guide
refdocs read https://example.test/some/guide --json
# Check the configuration without making a request
refdocs check
refdocs --versionExit codes: 0 on success, 1 when a lookup fails or the
configuration is not ready (the error names the problem), 2 for
usage errors.
MCP server
One bin serves both faces. The mcp subcommand starts a stdio MCP
server:
npx -y @reneturcios/refdocs mcpMCP client configuration:
{
"mcpServers": {
"refdocs": {
"command": "npx",
"args": ["-y", "@reneturcios/refdocs", "mcp"],
"env": {
"REF_API_KEY": "your-key-here"
}
}
}
}Tools
Tool | Input | What it returns |
|
| Matching documentation results from public documentation and GitHub, each with its title, URL, and content or snippet when the upstream provides them. |
|
| The page content converted to markdown: headings, links, lists, emphasis, and code blocks are kept in markdown form, while scripts, styles, and other non-content markup are dropped. |
Search responses are parsed tolerantly: a bare array, a results
list, and a documents list are all accepted. Reading plain-text
URLs returns the text unchanged.
Development
npm install
npm run build # tsc -> dist/
npm test # node test runner over test/*.test.ts, no network
npm run dev # run the CLI from source via tsxTests run entirely against a fake fetch function; no live API calls are made.
License
MIT — see LICENSE.
This server cannot be deployed
Maintenance
Related MCP Connectors
Web search, fetch, extract, and research for AI agents. Markdown output + AI-synthesized answers.
Web data for agents. Extract web pages, crawl websites, discover URLs, and search the web.
Web data for agents: YouTube transcripts, screenshots, Google News, WHOIS, jobs, tech stack, more.
Give your agent live data from Twitter, Reddit, the web and GitHub. No API keys, no scraping stack.
Related MCP Servers
- FlicenseNot gradedqualityDmaintenanceEnables AI agents to perform web searches, extract webpage content, and conduct end-to-end search-and-extract operations using multiple search providers and content extraction methods.-
- AlicenseAqualityAmaintenanceEnables AI agents to search and fetch high-quality information from multiple sources, including general web APIs, Wikipedia, arXiv, Hacker News, Stack Exchange, and Crossref, with optional pro-mode deep research and clean markdown page extraction.10MIT
- AlicenseNot gradedqualityAmaintenanceEnables AI agents to perform live web searches with ranked results and extracted page passages, plus clean markdown extraction of any URL.MIT
- AlicenseAqualityBmaintenanceEnables agents to search the live web and fetch URLs as clean Markdown, returning ranked search results and paginated page content through two lightweight tools.225 npmMIT