mcp-installation.mdx

title	MCP Installation
description	Connect Insforge to your AI coding assistant using Model Context Protocol

What is MCP?

Model Context Protocol (MCP) is an open standard that allows AI assistants to interact with external tools and databases. Insforge provides an MCP server that gives your AI direct access to:

• Database Management: Create/modify/delete tables, view schemas, manage table relationships
• Storage Buckets: Create/list/delete storage buckets for file uploads
• Backend Metadata: Get API keys, backend configuration, and system information
• SDK Documentation: Access SDK usage instructions for database, auth, and storage, AI Integration operations

Quick Installation

The easiest way to install MCP for your AI tool:

<Steps>
  <Step title="Create or select your project">
    Visit [https://insforge.dev](https://insforge.dev) and either:
    - Create a new project, or
    - Select an existing project
  </Step>
  
  <Step title="Click Connect">
    Click the **Connect** button to integrate with Agents
  </Step>
  
  <Step title="Install the MCP">
    <img src="/images/dashboard-connect-expanded.png" alt="Insforge Connect Dialog" />
    
    Select your AI tool tab and use the **Copy** button to copy the command, then run it in your terminal.
    
    The command includes your API key and base URL automatically:
    ```bash
    npx @insforge/install --client cursor \
      --env API_KEY=ik_d02a35cfd8056c18e9e59b34bf8bf773 \
      --env API_BASE_URL=https://trqnn5z3.us-east.insforge.app
    ```
  </Step>
</Steps>

<Note>
  The Copy button provides a ready-to-run command with your API key and base URL already configured for your specific project
</Note>

```bash npx @insforge/install --client [your-tool] \ --env API_KEY=your-api-key \ --env API_BASE_URL=http://localhost:7130 ```

<Note>
  Get your API key from your local dashboard at `http://localhost:7131`
</Note>

Replace [your-tool] with one of: cursor, claude, windsurf, cline, roo-code

Manual Installation

For tools that support direct MCP configuration:

Step 1: Install the MCP server

npm install -g @insforge/mcp-server

Step 2: Configure your AI tool

Edit your Claude configuration file:

**macOS**: `~/Library/Application Support/Claude/claude_desktop_config.json`
**Windows**: `%APPDATA%\Claude\claude_desktop_config.json`

```json
{
  "mcpServers": {
    "insforge": {
      "command": "npx",
      "args": ["-y", "@insforge/mcp-server"],
      "env": {
        "API_KEY": "your-api-key",
        "API_BASE_URL": "https://your-app.us-east.insforge.app"
      }
    }
  }
}
```

1. Install the Cline extension in VS Code 2. Open Cline settings (`Cmd/Ctrl + Shift + P` → "Cline: Open Settings") 3. Go to MCP Servers section 4. Add new server:

```json
{
  "insforge": {
    "command": "npx",
    "args": ["-y", "@insforge/mcp-server"],
    "env": {
      "API_KEY": "your-api-key",
      "API_BASE_URL": "https://your-app.us-east.insforge.app"
    }
  }
}
```

Cursor uses a composer file. Run the installer:

```bash
npx @insforge/install --client cursor \
  --env API_KEY=your-api-key \
  --env API_BASE_URL=https://your-app.us-east.insforge.app
```

Or manually add to `.cursorrules` in your project:

```
You have access to Insforge backend at https://your-app.us-east.insforge.app
Use the MCP tools to interact with the database, auth, and storage.
```

Verify Installation

Test your MCP connection by asking your AI assistant:

"Using Insforge MCP, show me what database tables exist"

Or for a more comprehensive test:

"Create a new table called 'test_mcp' with id and name fields"

Your AI should be able to:

• List existing tables
• Create new tables
• Query data
• Manage authentication
• Handle file uploads

Available MCP Tools

Once connected, your AI assistant has access to these tool categories:

Create, modify, and manage database tables and schemas Manage storage buckets for file uploads and organization Access backend configuration and system information Retrieve API documentation and implementation guides

Troubleshooting

1. Restart your AI tool after configuration 2. Check the configuration file path is correct 3. Verify your API key is valid 4. Ensure the API_BASE_URL is accessible - Verify your API key in the dashboard - Check that the API_BASE_URL matches your backend - For local development, ensure Docker is running - For cloud: Check your internet connection - For local: Verify Docker is running with `docker ps` - Test the backend directly: `curl [API_BASE_URL]/health`

Next Steps

Detailed setup for each AI tool Learn about database operations Build your first app Explore all available APIs