AI MCP Integration

MCP (Model Context Protocol) is an open protocol created by Anthropic that enables AI tools to connect to external systems in a standardized way. Microi MCP Server (Open Source Repository) allows AI tools to directly connect to the Microi platform, query database schemas in real-time, read engine code, and execute engines remotely.

What is Microi MCP

Microi MCP Server enables AI tools like GitHub Copilot, Cursor, and Claude Code to directly operate the Microi platform — no more manually copying and pasting table structures or API docs. AI can retrieve your database schema and business code in real-time.

AI Capabilities

Tool	Function	Read/Write
microi_get_status	Check backend connection status	Read-only
microi_get_db_schema	Get database schema (table names, fields, types, descriptions)	Read-only
microi_list_engines	List all API engines	Read-only
microi_get_engine_code	Get API engine JavaScript source code	Read-only
microi_save_engine_code	Save API engine code	Read-write
microi_create_engine	Create a new API engine	Read-write
microi_run_engine	Execute API engine remotely	Read-write
microi_list_events	List all V8 form events	Read-only
microi_get_event_code	Get V8 event source code	Read-only
microi_save_event_code	Save V8 event code	Read-write

Prerequisites

• Microi backend service deployed
• An AI programming tool installed (GitHub Copilot / Cursor / Claude Code)

Recommended: VS Code Extension (Zero Configuration)

Most users don't need manual MCP configuration

After installing the Microi VS Code Extension, MCP is automatically configured and ready to use out of the box.

After installing the extension, it automatically:

• Generates .vscode/mcp.json (GitHub Copilot) and .cursor/mcp.json (Cursor)
• Auto-refreshes tokens, no password storage needed
• Injects AI instruction files (.github/copilot-instructions.md, CLAUDE.md, .cursorrules)

Workflow: Install extension → Configure server connection → Pull code → MCP is ready.

The following content is for users who don't use the VS Code extension or need SSE remote deployment.

Manual Configuration: Local stdio Mode

The AI tool automatically launches the MCP Server process on each startup.

Installation

git clone https://gitee.com/microi-net/microi.mcp.git
cd microi.mcp
npm install
npm run build

GitHub Copilot (VS Code)

Add to .vscode/mcp.json in your project:

{
  "servers": {
    "microi": {
      "type": "stdio",
      "command": "node",
      "args": ["/path/to/microi.mcp/dist/index.js"],
      "env": {
        "MICROI_API_URL": "https://your-api-url",
        "MICROI_USERNAME": "username",
        "MICROI_PASSWORD": "password",
        "MICROI_OS_CLIENT": ""
      }
    }
  }
}

Cursor

Create .cursor/mcp.json in your project root:

{
  "mcpServers": {
    "microi": {
      "command": "node",
      "args": ["/path/to/microi.mcp/dist/index.js"],
      "env": {
        "MICROI_API_URL": "https://your-api-url",
        "MICROI_USERNAME": "username",
        "MICROI_PASSWORD": "password",
        "MICROI_OS_CLIENT": ""
      }
    }
  }
}

Claude Code (CLI)

claude mcp add microi -- \
  env MICROI_API_URL=https://your-api-url \
  env MICROI_USERNAME=username \
  env MICROI_PASSWORD=password \
  node /path/to/microi.mcp/dist/index.js

Security Notice

Configuration files contain sensitive information. Add .vscode/mcp.json and .cursor/mcp.json to .gitignore to prevent committing to Git.

Remote SSE Mode (Team / Production Recommended)

Deploy MCP Server as a Docker container so everyone connects to the same SSE endpoint.

Docker Deployment

cd microi.mcp
cp .env.example .env
# Edit .env with backend URL and admin credentials
docker compose up -d

Nginx Reverse Proxy

Recommended to mount under your existing API domain:

# MCP SSE endpoint
location /mcp/sse {
    proxy_pass http://127.0.0.1:3000/sse;
    proxy_http_version 1.1;
    proxy_set_header Connection '';
    proxy_buffering off;
    proxy_cache off;
    proxy_read_timeout 86400s;
}

# MCP message endpoint
location /mcp/messages {
    proxy_pass http://127.0.0.1:3000/messages;
    proxy_http_version 1.1;
}

# MCP health check
location /mcp/health {
    proxy_pass http://127.0.0.1:3000/health;
}

Verify Deployment

curl https://api.example.com/mcp/health
# Should return {"status":"ok","server":"microi-mcp-server","version":"1.0.0"}

Connect AI Tools via SSE

GitHub Copilot (.vscode/mcp.json):

{
  "servers": {
    "microi": {
      "url": "https://api.example.com/mcp/sse",
      "headers": {
        "X-Microi-Username": "username",
        "X-Microi-Password": "password",
        "X-Microi-OsClient": ""
      }
    }
  }
}

Cursor (.cursor/mcp.json):

{
  "mcpServers": {
    "microi": {
      "url": "https://api.example.com/mcp/sse",
      "headers": {
        "X-Microi-Username": "username",
        "X-Microi-Password": "password",
        "X-Microi-OsClient": ""
      }
    }
  }
}

Environment Variables

Variable	Required	Description	Example
MICROI_API_URL	✅	Microi backend API URL	https://api.microi.net
MICROI_USERNAME	※	Login username (required without Token)	admin
MICROI_PASSWORD	※	Login password (required without Token)	password
MICROI_OS_CLIENT	No	Application identifier (multi-tenant)	myApp
MICROI_TOKEN	※	Direct Token (takes priority over username/password)	Bearer xxx

Usage Examples

After configuring MCP, you can directly operate the platform in AI conversations:

You: Show me all tables in the current database
AI: (calls microi_get_db_schema) The current database has 42 tables...

You: Create an API engine to query the order list
AI: (calls microi_get_db_schema to understand schema → generates code → calls microi_create_engine)
    API engine created with Key: get-order-list...

You: Execute it and show me the results
AI: (calls microi_run_engine) Returned 20 order records...

Relationship with Skills and VS Code Extension

Solution	Capabilities	Use Case
VS Code Extension	V8 API knowledge + database schema + code push/pull + debugging	Daily development
MCP Server (this document)	Real-time database query, read/save engine code, remote execution	AI real-time platform operations
Skills	Scenario-specific coding best practices and code templates	Coding standards, in-depth guidance

Recommended: Use all three together

VS Code Extension provides API knowledge and schema → MCP provides real-time data query and remote execution → Skills provide coding best practices.