Platoona MCP Server
An MCP server that connects AI assistants to 10,000+ SaaS integrations through Platoona Connect, enabling search, connection, and execution of integration actions via natural language.
README
Platoona MCP Server
An MCP (Model Context Protocol) server that connects AI assistants like Claude Code, Cursor, and others to 10,000+ SaaS integrations through Platoona Connect.
Features
- search-tools - Semantic search (RAG) for 10K+ integration actions
- list-integrations - Browse available integrations
- list-connections - View your active connections
- connect-app - Connect to integrations (OAuth or API key)
- execute-tool - Execute integration actions
- disconnect-app - Disconnect from integrations
Prerequisites
- A Platoona Connect account
- Your Platoona API key (get it from Platoona Connect Portal)
- Node.js 18+
Installation
Option 1: Install from npm (recommended)
npm install -g @platoona/mcp
Option 2: Install from source
git clone https://github.com/platoona/platoona-mcp.git
cd platoona-mcp
npm install
npm run build
Configuration
Environment Variables
| Variable | Required | Description |
|---|---|---|
PLATOONA_API_KEY |
Yes | Your Platoona Connect API key |
PLATOONA_BASE_URL |
No | Custom API URL (for self-hosted) |
Note: User ID is automatically derived from your API key using a deterministic hash. This ensures consistent identification across sessions - same API key always produces the same user ID (mcp_<hash>).
Usage with Claude Code
Add to your Claude Code configuration (~/.claude/claude_desktop_config.json):
{
"mcpServers": {
"platoona": {
"command": "npx",
"args": ["@platoona/mcp"],
"env": {
"PLATOONA_API_KEY": "platoona_xxx..."
}
}
}
}
Or if installed globally:
{
"mcpServers": {
"platoona": {
"command": "platoona-mcp",
"env": {
"PLATOONA_API_KEY": "platoona_xxx..."
}
}
}
}
Usage with Cursor
Add to your Cursor MCP settings:
{
"mcpServers": {
"platoona": {
"command": "npx",
"args": ["@platoona/mcp"],
"env": {
"PLATOONA_API_KEY": "platoona_xxx..."
}
}
}
}
Available Tools
search-tools
Search for integration tools using semantic/RAG search.
Input:
- query: "send a message to slack"
- integrationFilter: "slack" (optional)
- limit: 10 (optional)
Output:
- List of matching tools with IDs, descriptions, and similarity scores
list-integrations
List available integrations that can be connected.
Input:
- search: "email" (optional)
- limit: 20 (optional)
Output:
- List of integrations with IDs, names, and slugs
list-connections
List all active connections for your MCP user.
Input: (none required)
Output:
- List of connections with IDs, status, and expiration
connect-app
Connect to an integration.
Input:
- integration: "slack" or UUID
- apiKey: "..." (for API key integrations)
- scopes: ["read", "write"] (for OAuth)
Output:
- For API key: connectionId and status
- For OAuth: authUrl to complete authorization
execute-tool
Execute an integration action.
Input:
- action: "slack:send-message" or UUID
- parameters: { channel: "#general", text: "Hello!" }
- timeout: 30000 (optional)
Output:
- Execution result with output data
disconnect-app
Disconnect from an integration.
Input:
- integration: "slack" or UUID
Output:
- Success confirmation
Example Workflow
-
Search for tools:
"Search for tools to send a Slack message" -
Check connections:
"List my Slack connections" -
Connect if needed:
"Connect to Slack" (will return OAuth URL) -
Execute the tool:
"Send a Slack message to #general saying Hello!"
API Endpoints Used
This MCP uses the dedicated /mcp/* endpoints for simplified integration:
| Endpoint | Purpose |
|---|---|
POST /mcp/tools/search |
RAG search for integration tools |
POST /mcp/connect |
Connect to integration (OAuth or API key) |
POST /mcp/execute |
Execute an integration action |
GET /mcp/connections |
List active connections |
POST /mcp/disconnect |
Disconnect from integration |
GET /ipaas/integrations |
List available integrations |
Development
# Install dependencies
bun install
# Run in development mode
bun run dev
# Build for production
bun run build
# Start production server
bun start
License
MIT
Support
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。