MCP LaunchBox
An MCP server that enables AI assistants to search and query your LaunchBox game library across all platforms including Steam, GOG, and emulators. It allows users to check ownership for bundles, view game metadata, find duplicates, and analyze playtime statistics through natural language.
README
MCP LaunchBox
An MCP server that lets AI assistants query your LaunchBox game library. Search your entire game collection across all platforms (Steam, GOG, Epic, emulators, etc.) directly from your AI assistant.
Works with Claude Desktop, Claude Code, or any MCP-compatible client.
Use Cases
- Send a screenshot of a game bundle and check which ones you already own — fuzzy matching handles OCR errors and typos
- Quick lookups before buying a game on sale
- Browse your library by platform or name
- Check how much time you've spent on a game
- Find installed games you haven't played yet
- See what you're currently playing across platforms
Setup
Requires Node.js 18+.
1. Have a LaunchBox Installation
You'll need LaunchBox installed with your game data.
2. Configure Your MCP Client
Add the server to your MCP client's configuration. There are two ways to do this:
Option A: Run directly from GitHub (no install needed)
{
"mcpServers": {
"launchbox": {
"command": "npx",
"args": ["-y", "github:danjam/mcp-launchbox"],
"env": {
"LAUNCHBOX_DATA_PATH": "/path/to/LaunchBox"
}
}
}
}
Option B: Clone and run locally
git clone https://github.com/danjam/mcp-launchbox.git
cd mcp-launchbox
npm install
Then point your MCP client at the local build:
{
"mcpServers": {
"launchbox": {
"command": "node",
"args": ["/path/to/mcp-launchbox/dist/index.js"],
"env": {
"LAUNCHBOX_DATA_PATH": "/path/to/LaunchBox"
}
}
}
}
Replace /path/to/mcp-launchbox with the actual path where you cloned the project.
What You Can Do
Search Games
Search the game library by title. Uses fuzzy matching across title and series fields.
Each result includes: id, title, platform, installed status, play time, and a confidence score (0-1, higher is better). Use get_game_details for full metadata.
Tool: search_games
| Parameter | Type | Required | Description |
|---|---|---|---|
query |
string | Yes | Search text to match against titles |
platform |
string | No | Filter by platform name |
limit |
number | No | Max results to return (default 25) |
Check Library
Check which games from a list are already in the library. Designed for bundle duplicate checking — pass all the titles in one call instead of searching one at a time. Omit platform to check across all platforms (recommended for bundles). Returns matches with confidence scores and a summary of how many you own vs. how many are new.
Tool: check_library
| Parameter | Type | Required | Description |
|---|---|---|---|
games |
string[] | Yes | Array of game title strings to look up (1-100) |
platform |
string | No | Filter matches to a specific platform |
Get Game Details
Get full details for a specific game by its ID. Returns metadata (title, platform, developer, publisher, genre, release date, series, etc.) and play data (play count, play time, last played, date added, installed, completed, progress, personal rating).
Tool: get_game_details
| Parameter | Type | Required | Description |
|---|---|---|---|
id |
string | Yes | The game ID (UUID from search results) |
include_notes |
boolean | No | Include the Notes field (default false) |
List Platforms
List all platforms in the library with their game counts. No parameters needed.
Tool: list_platforms
Find Duplicates
Find games owned on multiple platforms, grouped by title.
Tool: find_duplicates
| Parameter | Type | Required | Description |
|---|---|---|---|
query |
string | No | Optional title filter (fuzzy match) |
limit |
number | No | Max duplicate groups to return (default 25) |
Get Stats
Get library summary statistics (total games, total platforms, top 10 platforms by game count). No parameters needed.
Tool: get_stats
Reload Library
Reload all game data from disk. Use after adding or removing games in LaunchBox. No parameters needed.
Tool: reload_library
License
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。