For developers

Dispatch MCP Server

Give Claude read access to Dispatch's published AI stories via the Model Context Protocol.

What it is

A read-only MCP server that exposes Dispatch's published stories, topics, and engagement stats. Connect it to Claude Desktop or any MCP-compatible client and ask Claude to summarise the latest AI news, find stories on a specific topic, or pull trust scores for research.

Only stories that cleared the Trust ≥ 50 threshold are available — the same filter that governs what publishes on the site.

Endpoint

https://deliver-ai.xyz/mcp

HTTP/Streamable-HTTP transport · no authentication required · read-only

Claude Code

Run this in any project to add the server:

claude mcp add --transport http dispatch https://deliver-ai.xyz/mcp

Or add the snippet above to a .mcp.json file in your project root.

Claude.ai (web)

Open Settings → Integrations → Add integration and paste the endpoint URL. No configuration file needed.

https://deliver-ai.xyz/mcp

Works in any Claude.ai plan that supports integrations.

Claude Desktop

Add this to your claude_desktop_config.json:

{
  "mcpServers": {
    "dispatch": {
      "type": "http",
      "url": "https://deliver-ai.xyz/mcp"
    }
  }
}

Restart Claude Desktop after saving. The server appears as dispatch in the tool list.

Available tools

search_storiestopic?, trustMin?, since?, keyword?
Search published stories by topic, minimum trust score, publish date, or keyword.
get_storyid
Fetch a single story by ID — full body, sources, and trust breakdown.
get_stats(none)
Aggregate stats: total stories, per-topic breakdown, average trust, and top stories by engagement.

Rate limits

60 requests per minute per IP. Exceeding the limit returns HTTP 429 with a Retry-After header. This is sufficient for any reasonable Claude session.