fpl-mcp
An MCP server for Fantasy Premier League that allows querying gameweek info, player recommendations, squad management, and executing transfers via FPL's public API with optional cookie authentication.
README
fpl-mcp
An MCP server for Fantasy Premier League. Reads use FPL's public API; authenticated calls use a PingOne OIDC refresh token.
Tools
| Tool | Description |
|---|---|
current_gameweek |
Current/next gameweek and deadline. |
recommend_players |
Rank players by form, value (points per £m), and upcoming fixture difficulty. Filters: position, max_cost, horizon, limit. |
my_squad |
Your current squad, bank, and free transfers. Needs auth. |
make_transfer |
Transfer a player out/in. Previews by default; set confirm=true to submit. Needs auth. |
update_fpl_credentials |
Replace the stored auth token from inside chat when it's expired or revoked — no restart needed. |
Setup
npm install
npm run build
cp .env.example .env # then fill in the values
Environment
| Var | Purpose |
|---|---|
FPL_ENTRY_ID |
Your team ID (the number in fantasy.premierleague.com/entry/<ID>/). Needed for my_squad and make_transfer. |
FPL_REFRESH_TOKEN |
PingOne OIDC refresh token — see .env.example for how to grab it from DevTools. Needed for my_squad and make_transfer. |
Read-only tools (current_gameweek, recommend_players) work without any env vars.
Auth notes
FPL retired the old users.premierleague.com cookie login in favor of PingOne OIDC (account.premierleague.com). The server:
- Exchanges
FPL_REFRESH_TOKENfor a short-lived access token and sends it asX-API-Authorization: Bearer <token>. - Refreshes proactively before expiry and on any 401/403.
- PingOne rotates the refresh token on every exchange — the server rewrites
FPL_REFRESH_TOKENin.enveach time so the new one survives a restart. - If the token is ever revoked or fully expired, call the
update_fpl_credentialstool from chat with a freshly copiedoidc.uservalue instead of editing.envby hand.
Run
node dist/server.js # speaks MCP over stdio
Client config example
{
"mcpServers": {
"fpl": {
"command": "node",
"args": ["/absolute/path/to/fpl-mcp/dist/server.js"],
"env": {
"FPL_ENTRY_ID": "1234567",
"FPL_REFRESH_TOKEN": "eyJ..."
}
}
}
}
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。