@flymyai/mcp-server

MCP (Model Context Protocol) server for FlyMyAI image generation with batch support.

Features

• Single Image Generation - Generate images from text prompts
• Batch Processing - Generate multiple images in parallel with configurable concurrency
• Automatic Retries - Built-in retry logic with exponential backoff
• Rate Limit Handling - Respects API rate limits with automatic backoff
• Comprehensive Logging - Configurable log levels for debugging
• Highly Configurable - Environment variables for all settings

Installation

npm install -g @flymyai/mcp-server
# or
npx @flymyai/mcp-server

Quick Start

1. Get your API key from FlyMyAI

2. Set the environment variable:

export FLYMYAI_API_KEY=your-api-key

3. Run the server:

flymyai-mcp

Configuration with MCP Clients

opencode

Add to your opencode.jsonc:

{
  "mcp": {
    "flymyai": {
      "type": "local",
      "command": ["npx", "@flymyai/mcp-server"],
      "environment": {
        "FLYMYAI_API_KEY": "your-api-key"
      }
    }
  }
}

Claude Desktop

Add to your Claude Desktop config (~/Library/Application Support/Claude/claude_desktop_config.json):

{
  "mcpServers": {
    "flymyai": {
      "command": "npx",
      "args": ["@flymyai/mcp-server"],
      "env": {
        "FLYMYAI_API_KEY": "your-api-key"
      }
    }
  }
}

Available Tools

flymyai_generate

Generate a single image from a text prompt.

Parameters: | Name | Type | Required | Default | Description | |------|------|----------|---------|-------------| | prompt | string | ✅ | - | Text description of the image | | model | string | | flymyai/flux-schnell | Model ID | | width | number | | 1024 | Image width (64-2048) | | height | number | | 1024 | Image height (64-2048) | | num_inference_steps | number | | - | Denoising steps (1-100) | | guidance_scale | number | | - | Prompt adherence (0-20) | | seed | number | | - | Random seed for reproducibility |

Examples:

Generate an image of a sunset over mountains with dramatic clouds
a cyberpunk city @dev
@schnell a quick sketch of a cat
realistic photo of mountains @hq

Model shortcuts in prompt: | Shortcut | Model | |----------|-------| | @schnell, @fast | flymyai/flux-schnell | | @dev, @hq, @quality | flymyai/flux-dev | | @sdxl, @lora | flymyai/sdxl-lora |

flymyai_batch

Generate multiple images in parallel.

Parameters: | Name | Type | Required | Description | |------|------|----------|-------------| | prompts | array/string | ✅ | Array of prompts or newline-separated string | | model | string | | Model ID for all prompts |

Input formats:

// Array of strings
["a red car", "a blue boat", "a green plane"]

// Array of objects with custom params
[
  {"prompt": "a cat", "params": {"width": 512}},
  {"prompt": "a dog", "params": {"seed": 42}}
]

// Newline-separated string
"a cat\na dog\na bird"

flymyai_models

List available image generation models.

Available Models

| Model ID | Name | Speed | Quality | |----------|------|-------|---------| | flymyai/flux-schnell | FLUX Schnell | Fast | Good | | flymyai/flux-dev | FLUX Dev | Slow | High | | flymyai/sdxl-lora | SDXL LoRA | Medium | High |

Environment Variables

| Variable | Required | Default | Description | |----------|----------|---------|-------------| | FLYMYAI_API_KEY | ✅ | - | Your FlyMyAI API key | | FLYMYAI_BASE_URL | | https://api.flymy.ai/api/v1 | API base URL | | FLYMYAI_DEFAULT_MODEL | | flymyai/flux-schnell | Default model for generation | | FLYMYAI_MAX_CONCURRENCY | | 5 | Max parallel requests for batch | | FLYMYAI_MAX_RETRIES | | 3 | Max retry attempts | | FLYMYAI_RETRY_DELAY_MS | | 1000 | Initial retry delay (ms) | | FLYMYAI_TIMEOUT_MS | | 120000 | Request timeout (ms) | | FLYMYAI_LOG_LEVEL | | info | Log level: debug, info, warn, error, silent |

Programmatic Usage

You can also use the client directly in your Node.js applications:

import { FlyMyAIClient, loadConfig } from '@flymyai/mcp-server'

const config = loadConfig()
const client = new FlyMyAIClient(config)

// Single image
const result = await client.generate({
  prompt: "a beautiful sunset",
  width: 1024,
  height: 1024,
})
console.log(result.image) // base64 string

// Batch generation
const results = await client.generateBatch([
  { id: "1", prompt: "a cat" },
  { id: "2", prompt: "a dog" },
  { id: "3", prompt: "a bird" },
])

Error Handling

The server includes comprehensive error handling:

• ValidationError: Invalid input parameters
• APIError: FlyMyAI API errors (with automatic retry for 5xx errors)
• TimeoutError: Request timeout (retryable)
• RateLimitError: Rate limit exceeded (automatic backoff)

Development

# Clone the repository
git clone https://github.com/FlyMyAI/flymyai-mcp-server.git
cd flymyai-mcp-server

# Install dependencies
npm install

# Build
npm run build

# Run locally
FLYMYAI_API_KEY=your-key npm start

# Watch mode
npm run dev

License

MIT © FlyMyAI