<picture> <source srcset="./banner-dark.png" media="(prefers-color-scheme: dark)"> <source srcset="./banner.png" media="(prefers-color-scheme: light)"> <img src="./banner-dark.png" alt="Frontal Banner"> </picture>

Frontal MCP Server

A standalone Model Context Protocol (MCP) server that provides seamless access to Frontal's cloud services (AI, Blob Storage, Functions, Graph Database, and Pipelines) through a standardized interface.

Quick Start

Installation

# Install globally
npm install -g @frontal/mcp-server

# Or install locally
npm install @frontal/mcp-server

Basic Setup

1. Get your API key from Frontal Dashboard

2. Set up environment:

export FRONTAL_API_KEY="your_api_key_here"

3. Start the server:

# For Claude Desktop (stdio transport)
frontal-mcp-server --transport stdio

# For web applications (HTTP transport)
frontal-mcp-server --transport http --port 3000

Claude Desktop Integration

Add to your Claude Desktop configuration:

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

{
  "mcpServers": {
    "frontal": {
      "command": "frontal-mcp-server",
      "args": ["--transport", "stdio"],
      "env": {
        "FRONTAL_API_KEY": "your_api_key_here"
      }
    }
  }
}

Features

• AI Services: Text generation, image creation, and embeddings
• Blob Storage: File upload, download, and management
• Functions: Serverless function execution and management
• Graph Database: Graph queries and node creation
• Pipelines: Data pipeline creation and execution
• Multi-Transport: Support for both stdio and HTTP
• Type Safe: Full TypeScript support with Zod validation
• Monitoring: Built-in metrics and logging

Available Tools

The Frontal MCP Server provides access to the following services:

AI Service

• ai-generate-text: Generate text using AI models
• ai-generate-image: Generate images from text prompts
• ai-embed: Generate text embeddings

Blob Storage

• blob-upload: Upload files to blob storage
• blob-list: List objects in storage buckets

Functions

• functions-invoke: Execute serverless functions
• functions-list: List deployed functions

Graph Database

• graph-query: Execute graph queries
• graph-create-node: Create graph nodes

Pipelines

• pipelines-create: Create data pipelines
• pipelines-run: Execute pipelines

See API Documentation for detailed tool specifications.

Configuration

Environment Variables

Variable	Description	Default	Required
FRONTAL_API_KEY	Frontal API key	-	Yes
FRONTAL_BASE_URL	API base URL	https://api.frontal.dev/v1	No
MCP_LOG_LEVEL	Log level	info	No
ENABLE_AI	Enable AI service	true	No
ENABLE_BLOB	Enable Blob service	true	No
ENABLE_FUNCTIONS	Enable Functions service	true	No
ENABLE_GRAPH	Enable Graph service	true	No
ENABLE_PIPELINES	Enable Pipelines service	true	No

CLI Options

frontal-mcp-server [options]

Options:
  -t, --transport <type>     Transport type (stdio|http) [default: "stdio"]
  -p, --port <number>        HTTP port (for http transport) [default: 3000]
  -h, --host <address>       HTTP host (for http transport) [default: "localhost"]
  -k, --api-key <key>        Frontal API key
  -c, --config <path>         Configuration file path
  -v, --verbose              Verbose logging
  --log-level <level>        Log level (error|warn|info|debug) [default: "info"]

Usage Examples

Basic Usage

# Start with stdio transport (for Claude Desktop)
FRONTAL_API_KEY=your_key ./dist/bin/frontal-mcp-server.js

# Start with HTTP transport for web integration
FRONTAL_API_KEY=your_key ./dist/bin/frontal-mcp-server.js --transport http --port 3000

# Enable only specific services
ENABLE_AI=true ENABLE_BLOB=false ./dist/bin/frontal-mcp-server.js

Programmatic Usage

import { FrontalMcpServer, createLogger } from '@frontal/mcp-server';

const config = {
  apiKey: 'your_api_key',
  transport: { transport: 'stdio' },
  services: {
    ai: true,
    blob: true,
    functions: false,
    graph: false,
    pipelines: false,
  }
};

const logger = createLogger({ level: 'info' });
const server = new FrontalMcpServer(config, logger);

await server.initialize();
await server.connectStdio();

Architecture

The MCP Server follows a modular adapter pattern:

1. Core Server: Handles MCP protocol and transport management
2. Service Adapters: Each Frontal service has its own adapter
3. Configuration: Centralized config with environment variable support
4. Transport Layer: Supports both stdio and HTTP transports

Development

Project Structure

mcp-server/
├── src/
│   ├── adapters/          # Service adapters for each Frontal service
│   ├── config/            # Configuration management
│   ├── server/            # Core MCP server implementation
│   ├── utils/             # Utilities (logging, etc.)
│   └── bin/               # CLI entry point
├── tests/                 # Test files
├── docs/                  # Documentation
└── examples/              # Integration examples

Scripts

# Build the project
bun run build

# Run tests
bun run test

# Run tests in watch mode
bun run test:watch

# Generate coverage report
bun run test:coverage

# Lint code
bun run lint

# Format code
bun run format

# Type check
bun run type-check

Current Status

Completed Features:

• Project structure and build system
• Core MCP server implementation
• Service adapters for all Frontal services
• Configuration management
• CLI interface with stdio transport
• Comprehensive test suite
• Complete API documentation

In Development:

• HTTP transport implementation
• Integration with real Frontal SDKs
• Advanced error handling
• Performance monitoring

Contributing

We welcome contributions! Please see our Developer Guide for detailed information.

1. Fork the repository
2. Create a feature branch
3. Make your changes
4. Run tests: bun run test
5. Submit a pull request

Documentation

• API Documentation - Complete API reference
• Usage Guide - Practical examples and integration patterns
• Developer Guide - Architecture and contribution guidelines

Troubleshooting

Common Issues

Server won't start:

# Check API key
echo $FRONTAL_API_KEY

# Validate configuration
frontal-mcp-server --validate-config

Connection issues:

# Test with different transport
frontal-mcp-server --transport http --port 3000

# Check logs
frontal-mcp-server --verbose

Performance issues:

# Enable debug logging
MCP_LOG_LEVEL=debug frontal-mcp-server

# Monitor resources
top -p $(pgrep frontal-mcp-server)

Getting Help

• GitHub Issues: Report bugs and request features
• Discord Community: Join our developer community
• Documentation: Check docs/ for detailed guides

License

MIT License - see LICENSE file for details.