Plinth HN MCP
A production-clean MCP server for the Hacker News API — top/new/best stories, items and users, with SRE-grade rate limiting, backoff, and structured logging.
Plinth HN MCP
A small, production-clean Model Context Protocol server that gives any MCP client (Claude Desktop, Claude Code, Cursor, ...) live Hacker News data — top/new/best story feeds, individual items, and public user profiles — straight from the free, no-API-key Hacker News Firebase API.
It's free and MIT-licensed. The point isn't just "Hacker News in your LLM" — it's a reference for what an MCP server should look like when you actually have to run it: rate-limited, retried, observable, and validated.
Why this one
Most example MCP servers are toys that fall over the first time an LLM calls a tool in a loop or the upstream hiccups. This one ships the boring stuff that keeps a server alive in production:
- Outbound rate limiting — a token bucket caps how fast we hit the upstream, which matters here because the story feeds fan out into many per-item fetches, so an over-eager agent can't get you throttled.
- Exponential backoff with full jitter — transient
429/5xx/network errors are retried with capped, jittered delays that honorRetry-After. - Structured JSON logging to stderr — one JSON object per line, ready for a log shipper. Never writes to stdout (that's the JSON-RPC channel — polluting it corrupts the session).
- Graceful shutdown — SIGTERM/SIGINT close the HTTP client cleanly.
- Input validation on every tool argument — bad ids, empty usernames, or out-of-range limits fail fast with a clear, model-recoverable error instead of a stack trace.
- A real health check tool — reports version and live upstream reachability.
- Optional API-key gate — off by default (the upstream is public); flip on with one env var if you expose this server to others.
Quickstart (one command)
With uv installed, no clone or install needed:
uvx plinth-hn-mcp
Or with pip:
pip install plinth-hn-mcp
plinth-hn-mcp
The server speaks MCP over stdio. It's meant to be launched by an MCP client, not run interactively — but the command above will start and block, logging a startup line to stderr, which confirms it's healthy.
Use with Claude Desktop / Claude Code
Add this to your MCP client config (for Claude Desktop:
claude_desktop_config.json; for Claude Code: claude mcp add or your
project's .mcp.json):
{
"mcpServers": {
"plinth-hn": {
"command": "uvx",
"args": ["plinth-hn-mcp"]
}
}
}
Then ask: "What are the top 5 stories on Hacker News right now, and who posted the top one?"
Tools
| Tool | What it does |
|---|---|
top_stories(limit=10) | Current front-page (top) stories. |
new_stories(limit=10) | Newest submitted stories. |
best_stories(limit=10) | Highest-scoring stories. |
get_item(id) | A single story/comment/job/poll by id. |
get_user(username) | Public user profile (karma, created, about). |
health_check() | Server version + live upstream reachability. |
Story summaries include id, title, url, score, by, comment count, and
an hn_url permalink. Every data tool returns a JSON object with ok: true and
the payload, or ok: false with error and message on failure.
Configuration
All optional. Defaults are tuned for the free public upstream.
| Env var | Default | Purpose |
|---|---|---|
PLINTH_API_KEY | (unset) | If set, every data tool requires a matching api_key arg. |
PLINTH_LOG_LEVEL | INFO | DEBUG / INFO / WARNING / ERROR. |
PLINTH_MAX_LIST_LIMIT | 50 | Max items a single list tool will hydrate. |
PLINTH_RATE_LIMIT_REQUESTS | 10 | Max upstream requests per window. |
PLINTH_RATE_LIMIT_WINDOW_SECONDS | 1.0 | Rate-limit window length. |
PLINTH_MAX_RETRIES | 4 | Retries on transient upstream failures. |
PLINTH_BACKOFF_BASE_SECONDS | 0.5 | Base for exponential backoff. |
PLINTH_BACKOFF_MAX_SECONDS | 20.0 | Backoff cap. |
PLINTH_REQUEST_TIMEOUT_SECONDS | 15.0 | Per-request timeout. |
Development
python -m venv .venv && source .venv/bin/activate
pip install -e ".[dev]"
pytest
Tests mock the network (via respx), so they pass fully offline.
Want the whole toolkit?
This repo is one polished example. If you're building MCP servers for real and want to skip the yak-shaving:
→ MCP Server Starter Kit
Want Python and TypeScript templates, more worked examples, Dockerfiles, and a deploy + directory-publishing guide? (how to actually get your server listed in the MCP registries). Same SRE hygiene as this repo, batteries included.
More from the brand at plinthkit.vercel.app.
License
MIT © Plinth. Data from the Hacker News API (public, no auth).
Related Servers
Alpha Vantage MCP Server
sponsorAccess financial market data: realtime & historical stock, ETF, options, forex, crypto, commodities, fundamentals, technical indicators, & more
Neo
sponsorNEO MCP lets Claude Code, Cursor and VS Code hand off complex AI engineering tasks like AI model evals, AI agent optimization and more to NEO.
SheetsData
Instant access to electronic component datasheets for AI agents — specs, pinouts, package info, and absolute max ratings extracted from manufacturer PDFs on demand.
AgentMemory Mesh MCP
Paid remote MCP for agent memory MCP, structured receipts, usage logs, and audit-ready evidence for agent and CI workflows.
Agent Monitor Relay MCP
Paid remote MCP for AI agent run monitoring, failure detection, tool-call incident replay, SLA receipts, and client status exports.
EOL MCP Server
Check software end-of-life (EOL) dates and support status using the endoflife.date API to provide accurate lifecycle and security information.
Bring Your AI
No-data MCP endpoint for Claude Code to Codex harness migration checks.
Gemini CLI
Integrates with the unofficial Google Gemini CLI, allowing file access within configured directories.
Pipelock
Firewall for AI agents. MCP proxy that scans tool calls for credential leaks, prompt injection, and tool description poisoning.
AI Pair Programmer (Ruby)
AI-powered tools for code review, brainstorming, performance analysis, and security review in Ruby.
ConfigCat
interacts with ConfigCat feature flag platform. Supports managing feature flags, configs, environments, products and organizations. Helps to integrate ConfigCat SDK, implement feature flags or remove zombie (stale) flags.
Loaditout
Universal registry for 20,000+ MCP servers and agent skills with A/B/C/F security grading, CLI installer (npx loaditout add), verified publisher badges, and curated packs.