Connect an AI agent to X with MCP

What the X MCP server provides

REST and MCP share the same API key, credit balance, request history, and billing.

• search_x searches public posts, accounts, and trends.
• get_x_user reads a public profile.
• get_x_user_timeline fetches posts, replies, or media.
• get_x_post reads details, threads, and retweeters.
• get_x_user_network pages through followers and following.

Add TwitterXAPI to your MCP client

Add one remote HTTP server. Replace the placeholder with your customer key and keep the configuration private.

{
  "mcpServers": {
    "twitterxapi": {
      "type": "http",
      "url": "https://twitterxapi.com/api/mcp",
      "headers": { "Authorization": "Bearer tx_live_..." }
    }
  }
}

Ask the agent for public X data

Ask for the task. A capable MCP client selects the right tool and supplies its arguments.

Find the latest public posts from @marclou and summarize the product updates.

Read this X thread and return each claim with its post ID.

Get every follower page for @marclou and save the usernames.

Use cursors in agent workflows

Collection tools return one X page per call. Tell the agent how far to continue, or let it stop when has_more becomes false. Set a cost boundary for open ended research.

Tool text uses compact Contentful JSON. Structured results keep stable IDs and cursors for follow up calls. Raw X responses and queued exports remain on REST.