steam-mcp
Connects Claude to your Steam library via the Steam Web API. Query your games, playtime, recently played history, and store metadata using natural language. Zero-config install via npx steam-mcp.
README
steam-mcp
A local Model Context Protocol (MCP) server that connects Claude to your Steam library. Ask natural-language questions about your games without ever touching the API yourself.
"What have I been playing lately?"
"Suggest something from my backlog I haven't touched yet."
"How many hours have I put into RPGs?"
"Tell me about Elden Ring."
Tools
| Tool | What it does |
|---|---|
get_library |
Your full game library — title, playtime, last played |
get_recently_played |
Games played in the last 2 weeks |
search_library |
Filter by name, playtime range, or unplayed status |
get_game_details |
Store metadata — genres, tags, Metacritic score, price |
refresh_library |
Clear the cache to fetch fresh data from Steam |
Requirements
- Node.js 18+
- A Steam account with a public profile and library
- A Steam Web API key (free, takes 30 seconds to get)
Setup
1. Get a Steam API key
Go to steamcommunity.com/dev/apikey, log in, and register a key. The domain field can be anything (e.g., localhost).
2. Find your Steam ID
You need either:
- Your vanity URL username — the custom part of your profile URL:
steamcommunity.com/id/YOUR_USERNAME - Your 64-bit Steam ID — find it at steamid.io if you don't have a vanity URL
3. Make your profile public
In Steam: your profile → Edit Profile → Privacy Settings → set My Profile and Game details to Public.
4. Run setup
npx steam-mcp setup
This validates your credentials and writes the Claude Desktop config automatically. Restart Claude Desktop when it's done.
<details> <summary>Manual config (if you prefer)</summary>
Find your Claude Desktop config file:
- macOS:
~/Library/Application Support/Claude/claude_desktop_config.json - Windows:
%APPDATA%\Claude\claude_desktop_config.json
Add to the mcpServers section:
{
"mcpServers": {
"steam-mcp": {
"command": "npx",
"args": ["-y", "steam-mcp"],
"env": {
"STEAM_API_KEY": "your_api_key_here",
"STEAM_ID": "your_vanity_url_or_64bit_id"
}
}
}
}
</details>
Local development
git clone https://github.com/jkiley129/steam-mcp.git
cd steam-mcp
npm install
cp .env.example .env
# Edit .env with your API key and Steam ID
npm run build
To use the local build with Claude Desktop, point the config at the built file:
{
"mcpServers": {
"steam": {
"command": "node",
"args": ["/absolute/path/to/steam-mcp/dist/index.js"],
"env": {
"STEAM_API_KEY": "your_api_key_here",
"STEAM_ID": "your_vanity_url_or_64bit_id"
}
}
}
}
Configuration
| Environment variable | Required | Description |
|---|---|---|
STEAM_API_KEY |
Yes | Your Steam Web API key |
STEAM_ID |
Yes | Your 64-bit Steam ID or vanity URL username |
STEAM_MCP_CACHE_DIR |
No | Override cache directory (default: ~/.steam-mcp/cache/) |
Caching
Steam data is cached locally to keep responses fast and avoid hitting rate limits:
| Data | Cache TTL |
|---|---|
| Library | 1 hour |
| Recently played | 30 minutes |
| Game details | 24 hours |
Use the refresh_library tool to clear the cache immediately (e.g., after buying new games).
Troubleshooting
"Steam returned no library data" Your Steam profile or game library is set to private. Go to Steam → your profile → Edit Profile → Privacy Settings and set both to Public.
"Could not resolve Steam vanity URL"
Check that STEAM_ID matches the username in your Steam profile URL (steamcommunity.com/id/YOUR_USERNAME). Alternatively, use your numeric 64-bit Steam ID.
"Could not validate Steam API key" Verify your API key at steamcommunity.com/dev/apikey. Keys are tied to your Steam account login.
Tools not showing in Claude
Restart Claude Desktop fully (quit from the system tray, not just close the window) after editing the config file. Check that Node.js 18+ is installed by running node --version in a terminal.
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。