Habr MCP server
habr-mcp
An MCP server for searching and reading Habr articles: converts an article's markup into readable markdown right in the agent's reply
Low risk
We rate an entry low when it mostly gives the agent instructions and reference material.
Why this level
- Only reads public articles and search results, no authorization required
Install
Manual install
pip install httpx mcp && python habr_mcp.pyCommand assembled from the code's dependencies and entry point; the repository has no separate install guide.
This is third-party code. Review the repository files before installing.
What it does
A FastMCP server calls Habr's internal API and gives the agent two actions: search articles by query with relevance sorting in Russian, and read a specific article in full. The article's HTML markup (headers, lists, code, links, bold, italics) is converted to markdown by its own small converter with no external parsing library. It is a single file, depending only on httpx and mcp.
Who it is for. For developers who want to search and read Habr articles from a conversation with an agent without opening the site.
Good fit when
- You want to quickly find Habr articles on a topic and get them as markdown
- You want an agent to read a specific article and reference it in its answer
Not a fit when
- You need article comments or author profile data: the server does not return these
- You need Habr's official documented API: this uses an undocumented internal endpoint
Example request
Find recent Habr articles about MCP servers and show me the content of the most relevant oneLimitations
It relies on Habr's undocumented internal habr.com/kek/v2 endpoint, which can change without notice. The HTML-to-markdown converter is minimal and may lose some formatting on unusual layouts. There is no README in the repository; purpose and usage are visible only from the code.
How to disable. Remove the habr server from your MCP client config.
MCP
- Transport
- stdio
- Authentication
- not required
Security check
- Only reads public articles and search results, no authorization required
README in short
The repository has no README: the single-file habr_mcp.py shows a FastMCP server with two tools, habr_search and an article reader, plus its own HTML-to-markdown converter supporting headers, lists, links, code and paragraphs.
FAQ
Do I need a Habr account?
No, the server uses public search and article reading without authorization.
Can it read non-Russian articles?
Search defaults to fl=ru and hl=ru; you would need to change those parameters in the code for another language.
Related
CLI, skill and Python library for browser control: the agent clicks, fills forms and reads pages over CDP
A skill that gathers the last 30 days of discussion on a topic from Reddit, X, YouTube, HN, Polymarket and GitHub into one brief
The Chrome DevTools team's official MCP server: the agent drives a live Chrome, reads network and console, and records performance traces
A fast Rust CLI for agent browser automation: accessibility snapshots with element refs, an MCP server and skills