sleeper-mcp

An MCP (Model Context Protocol) server for the Sleeper Fantasy Sports API. Gives Claude, Cursor, VS Code Copilot, and any other MCP-compatible AI client read access to Sleeper's public API — no API key required.

What you can do

Ask your AI assistant things like:

• "What leagues am I in this season?" (give it your Sleeper username)

• "Show me the standings and rosters for my league"

• "Who won the championship bracket?"

• "What trades happened this week?"

• "Which players are being added the most right now?"

• "Show me all first-round draft picks from my 2024 draft"

Related MCP server: Sleeper Fantasy MCP

Available Tools

Prerequisites

• Node.js 18+ — download — OR — Bun — download

• A Sleeper account is helpful but not required (the API is fully public)

Installation

git clone https://github.com/yooitsgreg/sleeper-mcp.git
cd sleeper-mcp

# pick your package manager:
npm install   # or: pnpm install / yarn / bun install

The install triggers a prepare hook that compiles TypeScript to dist/. No separate build step needed.

Configuration

Pick the section for your AI client.

Claude Desktop

Edit ~/Library/Application Support/Claude/claude_desktop_config.json (macOS) or %APPDATA%\Claude\claude_desktop_config.json (Windows):

{
  "mcpServers": {
    "sleeper": {
      "command": "node",
      "args": ["/absolute/path/to/sleeper-mcp/dist/index.js"]
    }
  }
}

Claude Code (CLI)

Edit ~/.claude/claude.json:

{
  "mcpServers": {
    "sleeper": {
      "command": "node",
      "args": ["/absolute/path/to/sleeper-mcp/dist/index.js"]
    }
  }
}

VS Code (GitHub Copilot / MCP extension)

Create or edit .vscode/mcp.json in your workspace:

{
  "servers": {
    "sleeper": {
      "type": "stdio",
      "command": "node",
      "args": ["/absolute/path/to/sleeper-mcp/dist/index.js"]
    }
  }
}

Cursor

Edit ~/.cursor/mcp.json:

{
  "mcpServers": {
    "sleeper": {
      "command": "node",
      "args": ["/absolute/path/to/sleeper-mcp/dist/index.js"]
    }
  }
}

After editing the config, restart your AI client.

Development

npm run dev     # watch mode (tsx), recompiles on save
npm run build   # compile TypeScript → dist/
npm run clean   # remove dist/

Project structure

src/
├── index.ts              # Entry point — wires up McpServer + StdioServerTransport
├── constants.ts          # API base URL, timeouts, character limits
├── types.ts              # TypeScript interfaces for all Sleeper API responses
├── services/
│   └── sleeperClient.ts  # Shared axios client and error handler
└── tools/
    ├── users.ts          # sleeper_get_user
    ├── leagues.ts        # All league tools (9 tools)
    ├── drafts.ts         # All draft tools (5 tools)
    ├── players.ts        # sleeper_get_nfl_players, sleeper_get_trending_players
    └── state.ts          # sleeper_get_nfl_state

Notes

• No authentication required — Sleeper's API is free and public.

• Rate limit — Sleeper asks that you stay under 1,000 requests/minute.

• sleeper_get_nfl_players fetches ~5 MB of data. Use the position_filter, team_filter, or search parameters to keep responses manageable. Sleeper recommends caching this data and only refreshing once per day.

• All tools are read-only — this server never writes to Sleeper.

License

MIT