About agentfeed
agentfeed is a content-extraction API for AI agents that turns RSS/Atom feeds into JSON and web pages into text or Markdown. It also offers the same pipeline through an MCP server and REST endpoints.
Key features
- 01Convert RSS/Atom feeds to JSON
- 02Extract page text and Markdown
- 03REST API and health endpoint
- 04Stateless, standard-library server
Use cases
- 01Retrieve feed entries for an agent
- 02Extract article content from a URL
- 03Connect MCP clients to feed and article tools
MCP support
The streamable-HTTP MCP server at `/mcp` exposes `fetch_feed` and `fetch_article`. MCP clients connect to the live endpoint and pass an API key in the `X-Api-Key` header.
Configuration
Local server config is the same for every client. Copy it into your client’s MCP config file.
We haven’t collected a package or remote endpoint for this server. Follow the instructions in the project README.
Tools
Local servers expose tools only once running on your machine, and the maintainer has not provided a tool list.
Details
From the project READMEAgentFeed — tiny content-extraction API for AI agents
One GET request turns any RSS/Atom feed into clean JSON, or any web page into
clean text/markdown. No SDK, no signup, no tracking. A live instance runs at
https://agentfeed.bigfloppy.net.
MCP (Model Context Protocol)
POST /mcp is a streamable-HTTP MCP server exposing the same pipeline as two tools: fetch_feed (RSS/Atom -> JSON) and fetch_article (page -> text/Markdown). Point any MCP client (Claude Desktop, IDEs, agent frameworks) at https://agentfeed.bigfloppy.net/mcp; pass your API key via the X-Api-Key header. Same quotas and pricing as the REST API.
Endpoints
GET /api/feed?url=<rss-or-atom-url> # feed -> {title, link, published, summary}[]
GET /api/article?url=<page-url> # page -> {title, author, text, markdown}
GET /api/health
curl 'https://agentfeed.bigfloppy.net/api/feed?url=https://hnrss.org/frontpage'
curl 'https://agentfeed.bigfloppy.net/api/article?url=https://blog.python.org/'
Features
- Agent-friendly JSON — flat, predictable shapes; summaries truncated to 500 chars, article text capped at 20k chars.
- Fair free tier — 30 requests/min per IP, no key needed. Paid key: $5 per 100k requests/year (card, Stripe checkout, or USDC on Base with on-chain self-verification — no email, no accounts).
- SSRF-safe outbound fetch — 2 MB / 15 s caps, gzip handled.
- Single process, stdlib server — SQLite-free, stateless; only a small JSON file for key state.
Run your own
pip install -r requirements.txt
python3 server.py # serves on :5000
Edit paid.py to set your own wallet address and Stripe codes, or ignore that
file's payment logic entirely — the API works without keys on the free tier.
License
MIT — see LICENSE.
Comments
New to MCP?
From what the protocol is to installing your first server, these answers cover most first questions.
An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.