The X Bookmarks MCP Server
The Tweetsmash MCP server gives AI assistants direct access to your X (Twitter) bookmarks over the Model Context Protocol. Run it locally with npx @tweetsmash/mcp or connect to the hosted endpoint at mcp.tweetsmash.com/api/mcp, and Claude, ChatGPT, Cursor, or Raycast can search, count, and label your saved posts. It is powered by the same token as the X bookmarks API.
Add it to Claude in one config block
Works with Claude Desktop, Claude Code, Cursor, Raycast, and any MCP-compatible client. Generate your token in the Tweetsmash dashboard, then paste this into your MCP config:
{
"mcpServers": {
"tweetsmash": {
"command": "npx",
"args": ["@tweetsmash/mcp"],
"env": { "TWEETSMASH_TOKEN": "your-token-here" }
}
}
}Prefer a remote server? Point your client at https://mcp.tweetsmash.com/api/mcp with your API token as a Bearer header. Step-by-step instructions for each client are in the MCP integration guide.
Seven tools for your bookmarks
Ask your agent things like: find the bookmarks about prompt engineering, label everything from last week, or how many bookmarks mention MCP. Semantic search understands meaning, not just keywords. No agent in your stack yet? Bookmarks also sync to Notion and Google Sheets automatically.
X bookmarks MCP: frequently asked questions
What is the Tweetsmash X bookmarks MCP server?
An MCP (Model Context Protocol) server that gives AI assistants direct access to your X (Twitter) bookmarks. Run it locally with npx @tweetsmash/mcp or use the hosted endpoint at mcp.tweetsmash.com/api/mcp, and your agent can search, count, and label your saved posts.
How do I connect Claude to my X bookmarks?
Generate an API token in your Tweetsmash dashboard, then add the tweetsmash server to your Claude Desktop or Claude Code MCP config with npx @tweetsmash/mcp and the token as an environment variable. Claude can then search and organize your bookmarks in chat.
What tools does the X bookmarks MCP server expose?
list_bookmarks, search_bookmarks with keyword and semantic search, get_bookmark_count, list_labels, add_labels_to_tweets, remove_labels_from_tweets, and setup_vector_store for semantic search indexing.
Is there a hosted MCP endpoint?
Yes. Connect any remote MCP client to https://mcp.tweetsmash.com/api/mcp with your Tweetsmash API token as a Bearer header, no local install needed. The server is also listed on Smithery.
Does it work with ChatGPT, Cursor, and Raycast?
Yes. Any MCP-compatible client works: Claude Desktop, Claude Code, ChatGPT, Cursor, Raycast, and others. The setup guide covers Claude Desktop, Raycast, Smithery, and manual configuration.
How is this different from the official X MCP server?
X's hosted MCP at api.x.com/mcp gives agents general platform access using your X developer credentials. The Tweetsmash MCP server is built for bookmarks: it searches your imported library semantically, works with your labels, and needs only your Tweetsmash API token.
Build your information consumption workflow from Twitter Bookmarks
Import Bookmarks to TweetsmashSee our pricing
Built for curating & reading best of Twitter