Climate Feeds

What Climate Feeds MCP server does

Climate Feeds MCP server exposes three tools for working with climate and energy publishing feeds. list_feeds returns the curated catalog with each feed's ID, title, category, and source. Its filters support the climate and energy categories, as well as keyword matching.

read_feed accepts an ID returned by list_feeds and produces normalized entries containing a title, link, publication date, and summary. An optional keyword filter can narrow the returned items. fetch_feed works with a feed URL directly and normalizes RSS, Atom, and RDF responses, making it suitable for sources outside the curated catalog.

How it works

The hosted MCP endpoint is available at https://gateway.pipeworx.io/climate-feeds/mcp. A client can use MCP discovery and tool calls against that URL. The endpoint also exposes shared Pipeworx meta-tools in addition to the climate-feed tools, so its tools/list response contains more tools than the three described here. The initialization response is the authoritative source for the current scope.

The hosted endpoint can route requests through the wider Pipeworx catalog with ask_pipeworx, and the full gateway is available separately at https://gateway.pipeworx.io/mcp. Both gateway URLs provide access to the same underlying data sources; the difference is which pack's tools appear directly in the listing.

Climate Feeds MCP server can also run locally through the npm package @pipeworx/mcp-climate-feeds. In standalone mode, it communicates over stdin and stdout and exposes only the pack's own tools, without the gateway meta-tools or gateway routing.

Setup and configuration

For a hosted connection, add the following MCP server entry to a compatible client:

{
  "mcpServers": {
    "climate-feeds": {
      "url": "https://gateway.pipeworx.io/climate-feeds/mcp"
    }
  }
}

For local stdio execution, configure the client to run npx -y @pipeworx/mcp-climate-feeds:

{
  "mcpServers": {
    "climate-feeds": {
      "command": "npx",
      "args": ["-y", "@pipeworx/mcp-climate-feeds"]
    }
  }
}

The package can also be started directly with npx -y @pipeworx/mcp-climate-feeds to confirm that it launches. HTTP clients that do not use MCP can call the gateway's tool endpoint with JSON, such as POST https://gateway.pipeworx.io/v1/tools/climate_feeds_list_feeds.

Tools and capabilities

• Discover curated climate and energy feeds.
• Filter feed listings by category or keyword.
• Read normalized entries from a selected curated feed.
• Filter feed items by keyword.
• Fetch and normalize RSS, Atom, or RDF feeds by URL.
• Use a direct fetch with a proxy fallback when a source blocks the gateway.

A typical workflow starts with list_feeds, selects an ID, and then calls read_feed. Use fetch_feed when the required source is not in the curated list.

Limitations and notes

The curated tools require the feed ID returned by list_feeds; read_feed is not described as accepting an arbitrary URL. Feed results depend on the source content and availability. For blocked sources, fetch_feed can fall back to a proxy, but the material does not specify guarantees about every feed.

Hosted pack connections include shared Pipeworx schemas, which increases the tool count shown by tools/list. Local execution avoids those shared tools but also does not provide ask_pipeworx routing. The full gateway and the pack endpoint use the same Pipeworx gateway infrastructure, while standalone mode does not make a gateway round trip.