Blink MCP

Model Context Protocol server for Blink project management APIs. This server enables AI assistants to interact with Blink projects, databases, and edge functions.

Features

• SQL Database Operations: Execute SQL queries with automatic database creation
• Edge Function Management: List and deploy edge functions to projects
• Authentication: Secure access using Personal Access Tokens (PAT)
• Input Validation: Comprehensive validation using Zod schemas
• Error Handling: Detailed error messages and proper error codes

Installation

npx blink-mcp@latest

Configuration

Claude Desktop

Add this to your Claude Desktop configuration file:

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

{
  "mcpServers": {
    "blink-management": {
      "command": "npx",
      "args": ["blink-mcp@latest"],
      "env": {
        "BLINK_ACCESS_TOKEN": "your-personal-access-token-here"
      }
    }
  }
}

Other MCP Clients

For other MCP clients, run:

BLINK_ACCESS_TOKEN=your-token npx blink-mcp@latest

Getting Your Access Token

1. Go to Blink Dashboard
2. Navigate to your project settings
3. Generate a Personal Access Token (PAT)
4. Copy the token and use it in your configuration

Available Tools

1. list_projects

List all projects for the authenticated user.

Parameters: None

Example:

{
  "name": "list_projects"
}

2. get_project_details

Get detailed information about a specific project.

Parameters:

• project_id (string): The project identifier

Example:

{
  "name": "get_project_details",
  "arguments": {
    "project_id": "your-project-id"
  }
}

3. execute_sql

Execute SQL queries on project databases. Automatically creates database if it doesn't exist.

Parameters:

• project_id (string): The project identifier
• sql (string): SQL query to execute

Example:

{
  "name": "execute_sql",
  "arguments": {
    "project_id": "your-project-id",
    "sql": "CREATE TABLE users (id SERIAL PRIMARY KEY, name TEXT);"
  }
}

4. get_database_info

Get database information for a project.

Parameters:

• project_id (string): The project identifier

Example:

{
  "name": "get_database_info",
  "arguments": {
    "project_id": "your-project-id"
  }
}

5. create_database

Create a new database for a project.

Parameters:

• project_id (string): The project identifier

Example:

{
  "name": "create_database",
  "arguments": {
    "project_id": "your-project-id"
  }
}

6. delete_database

Delete the database for a project.

Parameters:

• project_id (string): The project identifier

Example:

{
  "name": "delete_database",
  "arguments": {
    "project_id": "your-project-id"
  }
}

7. list_edge_functions

List all edge functions deployed to a project.

Parameters:

• project_id (string): The project identifier

Example:

{
  "name": "list_edge_functions",
  "arguments": {
    "project_id": "your-project-id"
  }
}

8. deploy_edge_function 🆕 Supabase Compatible

Deploy an edge function to a project. 100% compatible with Supabase edge function format.

Parameters:

• project_id (string): The project identifier
• name (string): Function display name (will be converted to kebab-case for slug)
• files (array): Array of files to deploy, each with name and content
• entrypoint_path (string, optional): Entry point file path (default: "index.ts")
• import_map_path (string, optional): Import map file path (will automatically enable import map if provided)

Example:

{
  "name": "deploy_edge_function",
  "arguments": {
    "project_id": "your-project-id",
    "name": "Hello World",
    "entrypoint_path": "index.ts",
    "import_map_path": "import_map.json",
    "files": [
      {
        "name": "index.ts",
        "content": "import \"jsr:@supabase/functions-js/edge-runtime.d.ts\";\n\nDeno.serve(async (req: Request) => {\n  const data = {\n    message: \"Hello there!\"\n  };\n  \n  return new Response(JSON.stringify(data), {\n    headers: {\n      'Content-Type': 'application/json',\n      'Connection': 'keep-alive'\n    }\n  });\n});"
      }
    ]
  }
}

Notes:

• JWT verification is always disabled for deployed functions
• Import map is automatically enabled when import_map_path is provided
• Function names are automatically converted to kebab-case for URL slugs

9. get_function_logs

Retrieve logs for a specific function.

Note: This feature may not be available through the management API and is provided as a placeholder.

Parameters:

• project_id (string): The project identifier
• function_slug (string): Function name/slug
• hours_back (number, optional): Hours to look back (1-168, default: 24)
• log_level (string, optional): Log level filter ("all", "errors", "warnings", "info")
• limit (number, optional): Maximum number of logs (1-1000, default: 50)

Error Handling

The server provides detailed error messages for common issues:

• Authentication errors: Invalid or missing access token
• Project access: User doesn't have access to the specified project
• Validation errors: Invalid parameters or missing required fields
• API errors: Issues with the underlying Blink API

Development

To run from source:

git clone <repository>
cd blink-mcp
npm install
npm run build
npm start

License

MIT

Support

For issues and questions:

• GitHub Issues: Repository Issues
• Documentation: Blink API Documentation

Changelog

1.0.0

• Initial release
• Support for SQL database operations
• Support for edge function management
• Personal Access Token authentication
• Comprehensive input validation and error handling