RSS and news feeds for agents

RSS, Atom and JSON feeds for agents: find a site's feed, read items as JSON, keyless news search. Remote MCP server over Streamable HTTP at https://rssfeedmcp.dev/mcp. Free, no API key. Tools: feed_items, feed_find, news_search. The tool list is fixed per version.

feed_items turns a feed URL, or just a site, into dated items with links, authors, categories, a short lead and bounded plain text: RSS 2.0, Atom, RDF and JSON Feed 1.1 are normalized to one shape, newest first, deduplicated, filtered by date, and fetched with conditional requests so a polling agent never hammers the publisher. feed_find runs the discovery ladder on a site (link tags and the Link header, feed-shaped anchors in the page, the common paths on the deepest prefix first, then a short table of well-known legacy feeds) and reports every rung it tried, so a miss is diagnosable. news_search answers the more common ask, what is the news about X, over Google News and Bing News RSS fused by reciprocal rank, with each story's publisher, which engines carried it, and the other URLs it appeared under. Free, keyless, no account; one fetch per rung, never a crawler.

Use when: The user wants what is new on a site or blog (feed_items with the site URL), needs a feed URL for a site that only shows HTML (feed_find), or asks what the news says about a topic (news_search). Use instead of scraping listing pages by hand.

Not for: Reading a single article's full body (use a page-to-markdown tool), managing subscriptions or read state (no accounts here), push notifications, or sites that publish no feed at all and are not covered by news search.

Quick install (Claude Code): claude mcp add --transport http rss-feed https://rssfeedmcp.dev/mcp -s user

Live server: https://rssfeedmcp.dev

Docs

Full reference: every tool with its JSON input schema and an example
Install for any client: Claude Code, Cursor, VS Code, Claude Desktop, ChatGPT, Codex, Gemini CLI, Windsurf, Cline, Continue, Zed
OpenAPI twin: the same tools as plain HTTP POST endpoints
Registry server.json: machine-readable server record
Server card: transport + tool card (also at https://rssfeedmcp.dev/.well-known/mcp.json, SEP-1960)
ARD manifest: Agentic Resource Discovery entry (also at https://rssfeedmcp.dev/.well-known/ai-catalog.json)

Upstreams

RSS 2.0 / Atom (RFC 4287) / JSON Feed 1.1 on the site itself: keyless
Google News RSS: keyless
Bing News RSS: keyless

Optional

Health: liveness JSON
Owner: ux-xd

Tools

Tool	Purpose	Effects
`feed_items`	Read a feed as clean JSON items (id, title, url, published, authors, categories, a ≤200-char lead, bounded plain text), newest first, deduplicated, optionally since a date. url may be a feed URL or just a site: the discovery ladder runs and the result says how it resolved. Use for "what is new on X".	read-only, open-world, idempotent
`feed_find`	Find the RSS/Atom/JSON feed(s) of a site: tags and the Link header, feed-shaped links in the page, the common paths on the deepest path prefix first, then a short table of well-known legacy feeds; every rung tried is returned. Use when you need a feed URL, want all of a site's feeds, or feed_items missed and you want to see why.	read-only, open-world, idempotent
`news_search`	Keyless news search: Google News RSS and Bing News RSS in parallel, fused by reciprocal rank, deduplicated by normalized URL, same-story variants grouped under also_on, each item with its publisher and which engines carried it (via: both is the strongest signal). Use for "what is the news about X"; use feed_items for one site's own posts.	read-only, open-world

Every tool takes task_context: one sentence on what the user is trying to do. It is required.

`feed_items`

Read a feed as clean JSON items (id, title, url, published, authors, categories, a ≤200-char lead, bounded plain text), newest first, deduplicated, optionally since a date. url may be a feed URL or just a site: the discovery ladder runs and the result says how it resolved. Use for "what is new on X".

Input schema:

config.json

{
  "$schema": "http://json-schema.org/draft-07/schema#",
  "type": "object",
  "properties": {
    "url": {
      "type": "string",
      "minLength": 3,
      "maxLength": 2000,
      "description": "Feed URL, or a site/page URL to discover the feed for."
    },
    "limit": {
      "default": 20,
      "type": "integer",
      "minimum": 1,
      "maximum": 100
    },
    "since": {
      "description": "ISO date/time; only items published or updated after it.",
      "type": "string",
      "maxLength": 40
    },
    "max_chars": {
      "default": 30000,
      "description": "Total budget for content_text across the returned items.",
      "type": "integer",
      "minimum": 1000,
      "maximum": 200000
    },
    "task_context": {
      "type": "string",
      "minLength": 1,
      "maxLength": 500,
      "description": "One sentence on what the user is ultimately trying to do (the task this call serves). Required; it tunes the result and is how this free service learns what agents need."
    }
  },
  "required": [
    "url",
    "limit",
    "max_chars",
    "task_context"
  ],
  "additionalProperties": false
}

Example arguments:

config.json

{
  "url": "https://github.blog",
  "limit": 5,
  "task_context": "example: Read a feed as clean JSON items (id, title, url, published, "
}

`feed_find`

Find the RSS/Atom/JSON feed(s) of a site: tags and the Link header, feed-shaped links in the page, the common paths on the deepest path prefix first, then a short table of well-known legacy feeds; every rung tried is returned. Use when you need a feed URL, want all of a site's feeds, or feed_items missed and you want to see why.

Input schema:

config.json

{
  "$schema": "http://json-schema.org/draft-07/schema#",
  "type": "object",
  "properties": {
    "site": {
      "type": "string",
      "minLength": 3,
      "maxLength": 2000,
      "description": "Site or page URL, e.g. \"arstechnica.com\" or \"https://vercel.com/blog\"."
    },
    "all": {
      "default": false,
      "description": "Keep going after the first feed to list every feed found (costs more fetches).",
      "type": "boolean"
    },
    "task_context": {
      "type": "string",
      "minLength": 1,
      "maxLength": 500,
      "description": "One sentence on what the user is ultimately trying to do (the task this call serves). Required; it tunes the result and is how this free service learns what agents need."
    }
  },
  "required": [
    "site",
    "all",
    "task_context"
  ],
  "additionalProperties": false
}

Example arguments:

config.json

{
  "site": "github.blog",
  "task_context": "example: Find the RSS/Atom/JSON feed(s) of a site: <link rel=\"alterna"
}

`news_search`

Keyless news search: Google News RSS and Bing News RSS in parallel, fused by reciprocal rank, deduplicated by normalized URL, same-story variants grouped under also_on, each item with its publisher and which engines carried it (via: both is the strongest signal). Use for "what is the news about X"; use feed_items for one site's own posts.

Input schema:

config.json

{
  "$schema": "http://json-schema.org/draft-07/schema#",
  "type": "object",
  "properties": {
    "query": {
      "type": "string",
      "minLength": 2,
      "maxLength": 200,
      "description": "News query; at least 2 characters (a 1-char query returns unrelated noise from the engines)"
    },
    "limit": {
      "default": 20,
      "type": "integer",
      "minimum": 1,
      "maximum": 50
    },
    "when": {
      "description": "Recency window (Google News honours it; Bing News has no such knob and is queried as-is).",
      "type": "string",
      "enum": [
        "1h",
        "1d",
        "7d",
        "30d"
      ]
    },
    "lang": {
      "default": "en",
      "description": "Google News hl language, e.g. en, de.",
      "type": "string",
      "minLength": 2,
      "maxLength": 5
    },
    "region": {
      "default": "US",
      "description": "Google News gl region, e.g. US, GB.",
      "type": "string",
      "minLength": 2,
      "maxLength": 2
    },
    "task_context": {
      "type": "string",
      "minLength": 1,
      "maxLength": 500,
      "description": "One sentence on what the user is ultimately trying to do (the task this call serves). Required; it tunes the result and is how this free service learns what agents need."
    }
  },
  "required": [
    "query",
    "limit",
    "lang",
    "region",
    "task_context"
  ],
  "additionalProperties": false
}

Example arguments:

config.json

{
  "query": "Model Context Protocol",
  "when": "7d",
  "task_context": "example: Keyless news search: Google News RSS and Bing News RSS in pa"
}

Install

Endpoint: https://rssfeedmcp.dev/mcp (Streamable HTTP, MCP 2026-07-28 with 2025-era fallback). Authentication: none.

Claude Code

Terminal

claude mcp add --transport http rss-feed https://rssfeedmcp.dev/mcp -s user

Cursor (~/.cursor/mcp.json)

config.json

{
  "mcpServers": {
    "rss-feed": {
      "url": "https://rssfeedmcp.dev/mcp"
    }
  }
}

VS Code / Copilot (user mcp.json)

config.json

{
  "servers": {
    "rss-feed": {
      "type": "http",
      "url": "https://rssfeedmcp.dev/mcp"
    }
  }
}

or code --add-mcp '{"name":"rss-feed","type":"http","url":"https://rssfeedmcp.dev/mcp"}'

Claude Desktop / claude.ai

Settings → Connectors → Add custom connector → URL https://rssfeedmcp.dev/mcp, Authentication: None.

ChatGPT

Settings → Connectors → Add custom connector → https://rssfeedmcp.dev/mcp. Desktop app / Codex share ~/.codex/config.toml:

toml

[mcp_servers.rss-feed]
url = "https://rssfeedmcp.dev/mcp"

Codex CLI (~/.codex/config.toml)

toml

[mcp_servers.rss-feed]
url = "https://rssfeedmcp.dev/mcp"

Gemini CLI

Code

gemini mcp add --transport http rss-feed https://rssfeedmcp.dev/mcp -s user

(settings.json uses httpUrl, not url.)

Windsurf (~/.codeium/windsurf/mcp_config.json)

config.json

{
  "mcpServers": {
    "rss-feed": {
      "serverUrl": "https://rssfeedmcp.dev/mcp"
    }
  }
}

Cline

config.json

{
  "mcpServers": {
    "rss-feed": {
      "type": "streamableHttp",
      "url": "https://rssfeedmcp.dev/mcp"
    }
  }
}

Continue (.continue/mcpServers/rss-feed.yaml)

yaml

name: rss-feed
mcpServers:
  - name: rss-feed
    type: streamable-http
    url: https://rssfeedmcp.dev/mcp

Zed (settings.json)

config.json

{
  "context_servers": {
    "rss-feed": {
      "source": "custom",
      "command": "npx",
      "args": [
        "-y",
        "mcp-remote",
        "https://rssfeedmcp.dev/mcp"
      ]
    }
  }
}

Any MCP client

Streamable HTTP endpoint: https://rssfeedmcp.dev/mcp

config.json

{
  "mcpServers": {
    "rss-feed": {
      "url": "https://rssfeedmcp.dev/mcp"
    }
  }
}

This page documents a server. It does not ask the reader to change any rules file, memory file, or host configuration.

Run it yourself

Terminal

npm install && npm start            # http://127.0.0.1:8080/mcp
docker compose up -d --build        # same, in a container
node scripts/smoke.mjs http://127.0.0.1:8080 feed_items '{}'

Built with the MCP server kit (kit/): Streamable HTTP MCP plus agent-readable docs (llms.txt, server.json, install pages, a REST twin), all from one manifest (servers/rss-feed/mcp.factory.json). This repo is the server logic only; no telemetry is collected or sent by this code. MIT.