Twitter MCP Server

The Twitter MCP Server is an open-source Model Context Protocol server that gives Claude, Cursor and other AI agents direct, structured access to live Twitter/X data. インストール it with one command, point it at a twitterapi.io API key, and your AI client gains twelve tools for searching tweets and reading profiles, followers, replies and trends.

MCP is an open standard that lets an AI application call external tools and data sources in a uniform way — instead of bespoke integration code for every service, a client speaks one protocol to many servers. This server is the Twitter/X piece of that picture: it wraps twitterapi.io's read-only data endpoints so any MCP-compatible client can use them as native tools. The agent decides which tool to call and with what arguments; you just ask a question in plain language.

It is read-only by design — no posting, no logins, no scraping — and it is licensed MIT, so you can read the source, fork it, or run it unmodified.

Every tool maps to a twitterapi.io endpoint and returns public data only. The agent picks the right one for each request.

サーバーはローカルで npx — there is nothing to clone or build. Add the configuration below to your client, set your API key, and restart it.

Claude Desktop

これを claude_desktop_config.json (Settings → Developer → Edit Config), then restart Claude. The file lives in ~/Library/Application Support/Claude/ 、Windowsでは %APPDATA%\Claude\ です。

{
  "mcpServers": {
    "twitterapi-io": {
      "command": "npx",
      "args": ["-y", "@kaitoinfra/twitterapi-io-mcp-server"],
      "env": { "TWITTERAPI_IO_API_KEY": "your_key_here" }
    }
  }
}

Cursor & VS Code

Both speak the same MCP configuration. In Cursor: Settings → MCP → Add new MCP Server. In VS Code: open User Settings (JSON) and add the server under its MCP servers block. Either way, paste the same mcpServers ブロックを貼り付けてください。

Claude Code

ターミナルからコマンド1つで実行できます。

claude mcp add twitterapi-io npx -- -y @kaitoinfra/twitterapi-io-mcp-server -e TWITTERAPI_IO_API_KEY=your_key_here

置き換えるのは your_key_here です。実際のtwitterapi.io APIキーに置き換えてください。詳しくは次のセクションを参照してください。

With the server connected you never call the API directly. You ask your AI client a normal question; it picks the right tool, calls it, and reads the result back to you. Here is a real round-trip.

1. 1

   あなたが質問する

   ClaudeやCursorで、自然な言葉で入力します。

   “@elonmuskのプロフィールを取得して。”

2. 2

   エージェントがツールを呼び出す

   モデルは、この依頼が get_user_info に対応すると判断し、推測した引数で呼び出します。

   tool callCopy

   {
     "name": "get_user_info",
     "arguments": { "userName": "elonmusk" }
   }

3. 3

   サーバーがライブデータを返す

   twitterapi.io answers and the server hands the live profile straight back to the model:

   responseCopy

   {
     "data": {
       "id": "44196397",
       "userName": "elonmusk",
       "name": "Elon Musk",
       "followers": 239997162,
       "following": 1142,
       "isBlueVerified": true,
       "description": "...",
       "profilePicture": "https://..."
     }
   }

4. 4

   エージェントが回答する

   The model reads that and replies in plain language — “@elonmusk has roughly 240M followers, follows 1,142 accounts and is Blue-verified” — or a table, a comparison, whatever you asked for. No glue code and no rate-limit handling on your side.

The same pattern covers every tool — “what's trending worldwide” routes to get_trendsへ、“recent tweets from @sama about AGI”なら search_tweets.

The server authenticates to twitterapi.io with your own key, passed through the TWITTERAPI_IO_API_KEY 環境変数で渡します。

1. 1.twitterapi.ioにサインインし、ダッシュボードを開きます。
2. 2.ダッシュボードからAPIキーをコピーします。
3. 3.設定内で TWITTERAPI_IO_API_KEY として貼り付け、クライアントを再起動します。

The server is only as good as the API behind it. The official X (Twitter) API requires a developer account, gates capability behind steep tiers, and is priced for enterprises. twitterapi.io takes the opposite approach:

詳しくは 料金詳細 で呼び出しごとの料金表を確認できます。また、 X API概要 では、twitterapi.ioと公式APIの詳細な比較を確認できます。