Feed
A server for fetching and parsing RSS, Atom, and JSON feeds.
feed-mcp
Bring RSS feeds to Claude Desktop — Read news, blogs, and updates directly in your AI conversations.
What is feed-mcp?
feed-mcp is a Model Context Protocol (MCP) server that lets Claude Desktop read RSS, Atom, and JSON feeds. Think of it as a bridge that connects your favorite websites' RSS feeds to Claude, so you can ask questions about the latest articles, get summaries, and stay updated—all from within your Claude chat.
Why use it?
- 📰 Stay informed — Read the latest news and blog posts without leaving Claude
- 🎯 Get summaries — Ask Claude to summarize multiple articles across different feeds
- 🔍 Deep dive — Research topics by querying specific feeds or articles
- ⚡ Save time — No need to open multiple websites to stay current
Quick Start
Step 1: Add to Claude Desktop
Open your Claude Desktop configuration file and add feed-mcp with your favorite feeds:
macOS: ~/Library/Application Support/Claude/claude_desktop_config.json
Windows: %APPDATA%\Claude\claude_desktop_config.json
{
"mcpServers": {
"feed-mcp": {
"command": "docker",
"args": [
"run", "-i", "--rm",
"ghcr.io/richardwooding/feed-mcp:latest",
"run",
"https://techcrunch.com/feed/",
"https://www.theverge.com/rss/index.xml"
]
}
}
}
Step 2: Restart Claude Desktop
Restart Claude Desktop to load the new configuration.
Step 3: Start chatting!
Try asking Claude:
- "What are the latest tech news headlines?"
- "Summarize the top 5 articles from my feeds"
- "Are there any articles about AI today?"
Popular Feed Collections
Technology News
{
"mcpServers": {
"tech-news": {
"command": "docker",
"args": [
"run", "-i", "--rm",
"ghcr.io/richardwooding/feed-mcp:latest",
"run",
"https://techcrunch.com/feed/",
"https://www.theverge.com/rss/index.xml",
"https://www.wired.com/feed/rss",
"https://feeds.arstechnica.com/arstechnica/index"
]
}
}
}
Security & Privacy
{
"mcpServers": {
"security-news": {
"command": "docker",
"args": [
"run", "-i", "--rm",
"ghcr.io/richardwooding/feed-mcp:latest",
"run",
"https://krebsonsecurity.com/feed/",
"https://www.schneier.com/blog/atom.xml",
"https://www.bleepingcomputer.com/feed/"
]
}
}
}
Web Development
{
"mcpServers": {
"webdev-news": {
"command": "docker",
"args": [
"run", "-i", "--rm",
"ghcr.io/richardwooding/feed-mcp:latest",
"run",
"https://css-tricks.com/feed/",
"https://www.smashingmagazine.com/feed/",
"https://hacks.mozilla.org/feed/"
]
}
}
}
Podcasts
{
"mcpServers": {
"podcasts": {
"command": "docker",
"args": [
"run", "-i", "--rm",
"ghcr.io/richardwooding/feed-mcp:latest",
"run",
"https://feeds.npr.org/510282/podcast.xml",
"https://feeds.npr.org/381444908/podcast.xml"
]
}
}
}
Try asking Claude:
- "What are the latest podcast episodes?"
- "Summarize the most recent episode from NPR Politics"
- "Are there any episodes about climate change this week?"
Using Your RSS Reader Feeds
Already have feeds in Feedly, Inoreader, or another RSS reader? Export them as OPML and use with feed-mcp:
{
"mcpServers": {
"my-feeds": {
"command": "docker",
"args": [
"run", "-i", "--rm",
"-v", "/path/to/your/feeds.opml:/feeds.opml:ro",
"ghcr.io/richardwooding/feed-mcp:latest",
"run", "--opml", "/feeds.opml"
]
}
}
}
How to export OPML:
- Feedly: Settings → OPML → Export
- Inoreader: Preferences → Folders and Tags → Export OPML
- NewsBlur: Account → Import/Export → Export Stories
- The Old Reader: Settings → Import/Export → Export
How Claude Reads Feeds
When you ask Claude about your feeds, here's what happens:
- Browse first — Claude gets a list of article titles and metadata
- Read selectively — Claude only fetches full content for articles you ask about
- Smart caching — Articles are cached to avoid re-fetching
This two-pass approach keeps responses fast and prevents overwhelming your conversation.
Example Usage
You: "What's new in tech today?"
Claude will:
- Browse your tech feed titles
- Summarize the latest headlines
- Ask if you want details on specific articles
You: "Tell me more about the first article"
Claude will:
- Fetch the full content of that article
- Provide a detailed summary or answer your questions
Features
- 🌐 Multiple formats — RSS, Atom, and JSON feeds
- 📱 Import from readers — OPML support for easy migration
- 💾 Smart caching — Efficient feed retrieval with automatic updates
- ⚡ Fast & reliable — Built-in rate limiting and error handling
- 🔒 Secure — URL validation and private IP blocking
- 🐳 Easy deployment — Docker and Podman support
Advanced Features
For power users, feed-mcp includes:
- Dynamic feed management — Add/remove feeds at runtime
- MCP Resources — Advanced filtering and real-time subscriptions
- Intelligent prompts — Analyze trends, monitor keywords, generate reports
- Circuit breakers — Automatic handling of failing feeds
- Custom configuration — Rate limiting, retries, connection pooling
See docs/ADVANCED.md for details.
Alternative Installation Methods
Go Install
If you have Go installed:
go install github.com/richardwooding/feed-mcp@latest
feed-mcp run https://techcrunch.com/feed/
Then configure Claude Desktop:
{
"mcpServers": {
"feed-mcp": {
"command": "feed-mcp",
"args": ["run", "https://techcrunch.com/feed/"]
}
}
}
Podman
Prefer Podman over Docker? Just replace docker with podman:
{
"mcpServers": {
"feed-mcp": {
"command": "podman",
"args": [
"run", "-i", "--rm",
"ghcr.io/richardwooding/feed-mcp:latest",
"run",
"https://techcrunch.com/feed/"
]
}
}
}
Troubleshooting
"Claude hit the maximum length for this conversation"
If you see this error:
- You're fetching too many large articles at once
- Try asking Claude to browse titles first, then read specific articles
- The server automatically limits content to prevent this
Feed not updating
Feeds are cached for 10 minutes by default. If you need fresh data:
- Wait a few minutes and try again
- Restart Claude Desktop to clear the cache
Private/localhost feeds
By default, localhost and private IP feeds are blocked for security. To enable:
{
"mcpServers": {
"feed-mcp": {
"command": "docker",
"args": [
"run", "-i", "--rm",
"ghcr.io/richardwooding/feed-mcp:latest",
"run", "--allow-private-ips",
"http://localhost:8080/feed.xml"
]
}
}
}
Documentation
- ADVANCED.md — Dynamic feed management, MCP Resources, intelligent prompts
- ARCHITECTURE.md — Technical details, architecture, development guide
- CLAUDE.md — Instructions for Claude Code when working with this codebase
Contributing
Contributions are welcome! See the architecture docs for technical details and development guidelines.
License
MIT License — See LICENSE for details.
相关服务器
Bright Data
赞助Discover, extract, and interact with the web - one interface powering automated access across the public internet.
URnetwork
High quality VPN and Proxy connections
Scrapfly
Scrapfly MCP Server gives AI agents a simple, unified way to scrape live web data with built-in anti-bot handling.
Docs Fetch MCP Server
Fetch web page content with recursive exploration.
Documentation Crawler
Crawl websites to generate Markdown documentation and make it searchable through an MCP server.
DidItChange
MCP-native website monitoring with AI-powered change summaries and natural conversation control.
CrawlForge MCP
CrawlForge MCP is a production-ready MCP server with 18 web scraping tools for AI agents. It gives Claude, Cursor, and any MCP-compatible client the ability to fetch URLs, extract structured data with CSS/XPath selectors, run deep multi-step research, bypass anti-bot detection with TLS fingerprint randomization, process documents, monitor page changes, and more. Credit-based pricing with a free tier (1,000 credits/month, no credit card required).
Cloudflare Playwright
Control a browser for web automation tasks using Playwright on Cloudflare Workers.
Olostep MCP Server
A server for web scraping, Google searches, and website URL lookups using the Olostep API.
Crawl4AI RAG
Integrate web crawling and Retrieval-Augmented Generation (RAG) into AI agents and coding assistants.
BrowserAct
BrowserAct MCP Server is a standardized MCP service that lets MCP clients connect to the BrowserAct platform to discover and run browser automation workflows, access results/files and related storage, and trigger real-world actions via natural language.