/ ディレクトリ / プレイグラウンド / 12306-mcp
← 全サーバー ↗ GitHub
● コミュニティ Joooook ⚡ 即起動

12306-mcp

作者 Joooook · Joooook/12306-mcp

Query China's 12306 railway tickets via AI — search trains, check availability, and plan routes with natural language.

12306-mcp is an MCP server that connects AI assistants to China's official 12306 railway system. It supports ticket availability queries, train filtering, transit stop lookups, and multi-leg transfer planning. Runs via npx, Docker, or HTTP mode for flexible deployment.

▶ デモを見る 📦 インストール 💡 ユースケース

なぜ使うのか

主な機能

ライブデモ

実際の動作

12306-mcp-2.replay ▶ 準備完了
0/0

インストール

クライアントを選択

~/Library/Application Support/Claude/claude_desktop_config.json  · Windows: %APPDATA%\Claude\claude_desktop_config.json
{
  "mcpServers": {
    "12306-mcp-2": {
      "command": "TODO",
      "args": [
        "See README: https://github.com/Joooook/12306-mcp"
      ]
    }
  }
}

Claude Desktop → Settings → Developer → Edit Config を開く。保存後、アプリを再起動。

~/.cursor/mcp.json · .cursor/mcp.json
{
  "mcpServers": {
    "12306-mcp-2": {
      "command": "TODO",
      "args": [
        "See README: https://github.com/Joooook/12306-mcp"
      ]
    }
  }
}

Cursor は Claude Desktop と同じ mcpServers スキーマを使用。プロジェクト設定はグローバルより優先。

VS Code → Cline → MCP Servers → Edit
{
  "mcpServers": {
    "12306-mcp-2": {
      "command": "TODO",
      "args": [
        "See README: https://github.com/Joooook/12306-mcp"
      ]
    }
  }
}

Cline サイドバーの MCP Servers アイコンをクリックし、"Edit Configuration" を選択。

~/.codeium/windsurf/mcp_config.json
{
  "mcpServers": {
    "12306-mcp-2": {
      "command": "TODO",
      "args": [
        "See README: https://github.com/Joooook/12306-mcp"
      ]
    }
  }
}

Claude Desktop と同じ形式。Windsurf を再起動して反映。

~/.continue/config.json
{
  "mcpServers": [
    {
      "name": "12306-mcp-2",
      "command": "TODO",
      "args": [
        "See README: https://github.com/Joooook/12306-mcp"
      ]
    }
  ]
}

Continue はマップではなくサーバーオブジェクトの配列を使用。

~/.config/zed/settings.json
{
  "context_servers": {
    "12306-mcp-2": {
      "command": {
        "path": "TODO",
        "args": [
          "See README: https://github.com/Joooook/12306-mcp"
        ]
      }
    }
  }
}

context_servers に追加。保存時に Zed がホットリロード。

claude mcp add 12306-mcp-2 -- TODO 'See README: https://github.com/Joooook/12306-mcp'

ワンライナー。claude mcp list で確認、claude mcp remove で削除。

ユースケース

実用的な使い方： 12306-mcp

How to find the cheapest train between two cities with 12306-mcp

👤 Travelers planning trips within China ⏱ ~5 min beginner

使うタイミング： You need to compare train options between two Chinese cities and find the best deal.

前提条件
  • 12306-mcp server running — npx -y 12306-mcp
フロー
  1. Search for available trains
    Search for trains from Beijing to Shanghai tomorrow. Show me all available options with prices and seat types.✓ コピーしました
    → A table of trains with departure times, durations, seat classes, and prices
  2. Filter by preference
    Which of these has second-class seats available and arrives before 6pm?✓ コピーしました
    → Filtered list matching the criteria

結果： A clear comparison of train options to make a booking decision.

注意点
  • Station names must match 12306's database exactly — Use the station search tool first to confirm the correct station name

Plan a multi-leg train journey with transfers using 12306-mcp

👤 Travelers heading to cities without direct trains ⏱ ~10 min beginner

使うタイミング： No direct train exists between your origin and destination.

前提条件
  • 12306-mcp server running — npx -y 12306-mcp
フロー
  1. Query transfer options
    I need to get from Chengdu to Lhasa by train. Find me transfer routes with the shortest total travel time.✓ コピーしました
    → One or more transfer itineraries with intermediate stations and connection times
  2. Check availability on each leg
    Check ticket availability for each leg of the recommended route on March 15.✓ コピーしました
    → Seat availability for both legs of the journey

結果： A complete multi-leg itinerary with confirmed availability.

注意点
  • Tight transfer windows may cause missed connections — Ask Claude to filter for transfers with at least 2 hours between trains

組み合わせ

他のMCPと組み合わせて10倍の力を

12306-mcp-2 + fetch

Combine train search with web fetch to check weather at your destination before booking

Find me a train from Shanghai to Hangzhou this Saturday, and check the weekend weather forecast for Hangzhou.✓ コピーしました

ツール

このMCPが提供する機能

ツール入力呼び出すタイミングコスト
query_tickets from: str, to: str, date: str Search for available train tickets between two stations 1 API call
search_stations keyword: str Look up station names and codes by keyword 1 API call
query_transfer from: str, to: str, date: str Find transfer routes when no direct train exists 1 API call
get_train_route_stations train_no: str Get all stops along a specific train's route 1 API call

コストと制限

運用コスト

APIクォータ
Depends on 12306 public API rate limits — typically sufficient for personal use
呼び出しあたりのトークン
200–800 tokens per query
金額
Free — queries the public 12306 API
ヒント
Batch station lookups before running ticket queries to avoid redundant calls.

セキュリティ

権限、シークレット、影響範囲

認証情報の保管： No credentials required — uses public 12306 API endpoints
データ送信先： All calls go to 12306.cn official API endpoints

トラブルシューティング

よくあるエラーと対処法

Station name not found

Use the station search tool to find the exact name. Chinese station names must match the 12306 database — try both simplified Chinese and pinyin.

確認： Search for the station by keyword first
No trains found for the given date

12306 only sells tickets 15 days in advance. Ensure your date is within the booking window.

確認： Try a date within the next 15 days
Connection timeout

The 12306 API can be slow during peak booking times (e.g., Spring Festival). Retry after a few minutes.

確認： Test with a simple station search to confirm connectivity

代替案

12306-mcp 他との比較

代替案代わりに使う場面トレードオフ
mcp-server-12306You want a Python-based alternative with FastAPI backend and additional price query toolsMore features (price queries, HTTP streaming) but requires Python/uv setup

その他

リソース

📖 GitHub の公式 README を読む

🐙 オープンな issue を見る

🔍 400以上のMCPサーバーとSkillsを見る