Show HN: Readhn – AI-Native Hacker News MCP Server (Discover, Trust, Understand)

Posted by xodn348 7 hours ago

Counter2Comment2OpenOriginal

I felt frustrated finding high-signal discussions on HN, and I started this project to better understand how this community actually works.

That led me to build readhn, an MCP server that helps with three things:

- Discover: find relevant stories/comments by keyword, score, and time window

- Trust: identify credible voices using EigenTrust-style propagation from seed experts

- Understand: show why each result is ranked, with explicit signals instead of a black-box score

It includes 6 tools: discover_stories, search, find_experts, expert_brief, story_brief, and thread_analysis.

I also added readhn setup so AI agents can auto-configure it (Claude Code, Codex, Cursor, and others) after pip install.

I’d love feedback on:

1) whether these ranking signals match how you evaluate HN quality,

2) trust-model tradeoffs,

3) what would make this useful in your daily workflow.

If this is useful to you, starring the repo helps others discover it: https://github.com/xodn348/readhn

Comments

Comment by socialinteldev 4 hours ago

the trust propagation angle is interesting — most mcp servers just expose raw data, the meta-quality-signal layer is underbuilt. one thing i've noticed building an mcp server: the challenge is less discoverability (smithery + server-card.json helps) and more that agents don't know how to evaluate whether a data source is worth the cost. eigentrust for hn quality signals could be a useful primitive for that

Comment by xodn348 42 minutes ago

Thank you for your feedback. I agree with your opinion about ai agents and all the information they cast is not really credible. As a result, I choose to start with the expert or reliable people in the community to understand it better.