Steam-MCP
MCP server for Valve's Steam that enables querying game libraries, player profiles, achievements, friends, store listings, Workshop items, and current player counts. Provides 14 tools across 5 categories with a React dashboard and REST bridge.
README
<p align="center"> <img src="https://img.shields.io/badge/python-3.12+-blue?logo=python" alt="Python"> <img src="https://img.shields.io/badge/fastmcp-3.2+-purple" alt="FastMCP"> <img src="https://img.shields.io/badge/license-MIT-green" alt="License"> <img src="https://img.shields.io/github/v/release/sandraschi/steam-mcp" alt="Release"> <img src="https://img.shields.io/badge/status-active-brightgreen" alt="Status"> </p>
<p align="center"> <img src="https://img.shields.io/badge/Port_Backend-11020-blueviolet" alt="Backend Port"> <img src="https://img.shields.io/badge/Port_Frontend-11021-blueviolet" alt="Frontend Port"> <img src="https://img.shields.io/badge/MCP_HTTP-%2Fmcp-ff69b4" alt="MCP Path"> </p>
Steam-MCP
MCP server for Valve's Steam — query your game library, player profile, achievements, friends, store listings, Workshop items, and current player counts. 14 tools across 5 categories, with a React dashboard and REST bridge.
Features
- Profile — look up yourself, any player, friends list, resolve vanity URLs
- Library — owned games with playtime, recently played, detailed store info (no key needed)
- Stats — per-game achievements, global achievement percentages, current player counts, leaderboards
- Store & News — latest news per game, search the store by name
- Workshop — query published workshop items by app, sort by popularity/recency
No API key needed for: store details, player counts, news, achievement percentages, and store search. API key unlocks profile, library, friends, and Workshop.
Quick Start
# 1. Clone & install
git clone https://github.com/sandraschi/steam-mcp.git
cd steam-mcp
uv sync
# 2. Set your Steam Web API key (free)
# Get one: https://steamcommunity.com/dev/apikey
$env:STEAM_API_KEY = "XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX"
$env:STEAM_ID = "7656119xxxxxxxxxx" # your 64-bit Steam ID
# 3. Start the server
just serve
Server runs on http://localhost:11020. MCP endpoint at http://localhost:11020/mcp. REST bridge at /api/status, /api/tools, /api/tools/{name}/call.
Tools
Profile
| Tool | Description | Key |
|---|---|---|
get_own_profile |
Player summary for configured STEAM_ID | Yes |
get_player_summaries(steamids) |
Look up any player(s) by Steam ID | Yes |
get_friend_list(steamid) |
List friends for a user | Yes |
resolve_vanity_url(vanity_url) |
Resolve custom URL to 64-bit Steam ID | Yes |
Library
| Tool | Description | Key |
|---|---|---|
get_owned_games(steamid) |
Full game library with playtime | Yes |
get_recently_played_games(steamid) |
Recently played games | Yes |
get_app_details(app_id) |
Store page data (name, price, description, screenshots) | No |
Stats
| Tool | Description | Key |
|---|---|---|
get_player_achievements(steamid, app_id) |
Player's achievements for a game | Yes |
get_global_achievement_percentages(app_id) |
Global completion rates | No |
get_number_of_current_players(app_id) |
Concurrent players right now | No |
get_game_leaderboards(app_id) |
List leaderboards for a game | Yes |
Store & News
| Tool | Description | Key |
|---|---|---|
get_news_for_app(app_id) |
Latest news articles | No |
search_store(query) |
Search games by name | No |
Workshop
| Tool | Description | Key |
|---|---|---|
query_workshop_items(app_id, query) |
Query Workshop items | Yes |
Usage
As an MCP server (for AI assistants)
Connect any MCP-compatible client to the STDIO or HTTP endpoint:
// Claude Desktop, opencode, etc.
{
"mcpServers": {
"steam": {
"command": "uv",
"args": ["run", "steam-mcp"],
"env": {
"STEAM_API_KEY": "...",
"STEAM_ID": "..."
}
}
}
}
Or via HTTP SSE:
{
"mcpServers": {
"steam": {
"url": "http://localhost:11020/mcp"
}
}
}
Via REST API
# Status
curl http://localhost:11020/api/status
# List tools
curl http://localhost:11020/api/tools
# Call a tool
curl -X POST http://localhost:11020/api/tools/get_number_of_current_players/call \
-H "Content-Type: application/json" \
-d '{"arguments": {"app_id": 440}}'
Via React Dashboard
start.ps1 # launches backend + frontend
# or
just frontend-dev # frontend only on http://localhost:11021
Godot + Steam Publishing
This MCP handles accessing Steam data. To publish a Godot game on Steam:
- GodotSteam (godotsteam.com) — GDExtension that wraps Steamworks SDK as native Godot nodes. Handles achievements, leaderboards, lobbies, networking, Workshop, overlay, DRM, cloud saves.
- Steamworks Partner — Register at partner.steamgames.com. $100 Steam Direct fee per game, recoupable after $1K revenue. 70/30 revenue split.
- SteamPipe — Upload builds via
steamcmd.exe. Supports delta patching, beta branches, rollbacks.
Architecture
src/steam_mcp/
server.py # FastAPI + MCP HTTP composition
web.py # REST bridge
transport.py # STDIO/HTTP transport
config.py # Env-based settings
mcp/
registry.py # Shared FastMCP instance
tools/ # Tool implementations (14 tools, 5 files)
profile.py library.py stats.py store.py workshop.py
webapp/ # Vite React dashboard (TypeScript, Tailwind)
native/ # Tauri 2.0 scaffold (placeholder)
Development
uv sync # Install Python deps
npm --prefix webapp i # Install frontend deps
just serve # Dev server with hot reload
just lint # Ruff check
npm run build -w webapp # Build frontend
Ports
| Service | Port |
|---|---|
| Backend (FastAPI + MCP HTTP) | 11020 |
| Frontend (Vite dev) | 11021 |
| MCP Streamable HTTP | POST /mcp on 11020 |
License
MIT
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。