together-mcp-ai

MCP server for Together AI image generation. Supports text-to-image, image editing (Kontext), and video (Kling) via 29+ models.

Quick Start

Add to your MCP client config:

{
  "mcpServers": {
    "together-image-gen": {
      "command": "npx",
      "args": ["-y", "together-mcp-ai@latest"],
      "env": {
        "TOGETHER_API_KEY": "<YOUR_API_KEY>"
      },
      "autoApprove": ["generate_image", "list_models"]
    }
  }
}

Get your API key at api.together.xyz → API Keys → Create.

Tools

| Tool | Description | |------|-------------| | generate_image | Text-to-image, image editing (Kontext), or video (Kling) | | list_models | Show all available models with costs and aliases |

Models & Pricing

All prices verified from api.together.ai/models — March 2026.

Black Forest Labs (FLUX)

| Model ID | Price | Notes | |----------|-------|-------| | black-forest-labs/FLUX.1-schnell | $0.003 | Fast, cheap — default | | black-forest-labs/FLUX.1-pro | $0.050 | High quality production (v1) | | black-forest-labs/FLUX.1.1-pro | $0.040 | Best FLUX v1 text-to-image quality | | black-forest-labs/FLUX.1-krea-dev | $0.025 | Creative/artistic generation | | black-forest-labs/FLUX.1-kontext-pro | $0.040 | Image editing (requires image_url) | | black-forest-labs/FLUX.1-kontext-max | $0.080 | Max fidelity image editing | | black-forest-labs/FLUX.2-dev | $0.0154 | FLUX v2 dev/testing | | black-forest-labs/FLUX.2-pro | $0.030 | FLUX v2 production quality | | black-forest-labs/FLUX.2-flex | $0.030 | FLUX v2 flexible generation | | black-forest-labs/FLUX.2-max | $0.070 | FLUX v2 maximum quality |

Google

| Model ID | Price | Notes | |----------|-------|-------| | google/imagen-4.0-preview | $0.040 | Google's latest image gen | | google/imagen-4.0-fast | $0.020 | Fast Google image gen | | google/imagen-4.0-ultra | $0.060 | Highest quality Google gen | | google/gemini-flash-image-2.5 | $0.039 | Gemini multimodal image gen | | google/gemini-3-pro-image | $0.134 | Gemini Pro (most expensive) |

HiDream

| Model ID | Price | Notes | |----------|-------|-------| | HiDream-ai/HiDream-I1-Full | $0.009 | Full quality | | HiDream-ai/HiDream-I1-Dev | $0.0045 | Dev/testing | | HiDream-ai/HiDream-I1-Fast | $0.0032 | Fast, near-free |

ByteDance (Seedream)

| Model ID | Price | Notes | |----------|-------|-------| | ByteDance-Seed/Seedream-3.0 | $0.018 | Seedream v3 | | ByteDance-Seed/Seedream-4.0 | $0.030 | Seedream v4 latest |

Stability AI

| Model ID | Price | Notes | |----------|-------|-------| | stabilityai/stable-diffusion-3-medium | $0.0019 | SD3 medium, very cheap | | stabilityai/stable-diffusion-xl-base-1.0 | $0.0019 | SDXL, very cheap |

RunDiffusion

| Model ID | Price | Notes | |----------|-------|-------| | RunDiffusion/Juggernaut-pro-flux | $0.0049 | Quality FLUX fine-tune | | Rundiffusion/Juggernaut-Lightning-Flux | $0.0017 | Ultra-cheap, fast |

Others

| Model ID | Price | Notes | |----------|-------|-------| | Qwen/Qwen-Image | $0.0058 | Qwen image generation | | Lykon/DreamShaper | $0.0006 | Cheapest model available | | ideogram/ideogram-3.0 | $0.060 | Best at rendering text in images | | openai/gpt-image-1.5 | $0.034 | GPT-powered image gen | | wan-ai/wan-2.6-image | $0.030 | Wan AI generation |

Budget Tiers

| Tier | Price Range | Models | |------|-------------|--------| | Ultra-cheap | $0.0006 - $0.002 | DreamShaper, Juggernaut-Lightning, SDXL, SD3-Medium | | Cheap | $0.003 - $0.005 | FLUX.1-schnell, HiDream-Fast, Juggernaut-Pro, HiDream-Dev | | Mid | $0.015 - $0.030 | FLUX.2-dev, Seedream-3.0, FLUX.2-pro, FLUX.2-flex, Seedream-4.0 | | Premium | $0.030 - $0.060 | FLUX.1.1-pro, Kontext-Pro, Imagen-4.0, GPT-Image-1.5, Ideogram-3.0 | | High | $0.060 - $0.134 | FLUX.2-max, Kontext-Max, Imagen-4.0-ultra, Gemini-3-Pro |

Tip: Use FLUX.1-schnell ($0.003) for iteration, switch to FLUX.2-pro ($0.030) or FLUX.1.1-pro ($0.040) for final renders.

Parameters

| Parameter | Type | Default | Description | |-----------|------|---------|-------------| | prompt | string | required | Text description of the image | | model | string | flux-schnell | Model alias or full model ID | | width | number | 1024 | Width in pixels (64-2048) | | height | number | 1024 | Height in pixels (64-2048) | | aspect_ratio | string | — | 1:1 16:9 9:16 4:3 3:4 3:2 2:3 | | n | number | 1 | Number of images (1-4) | | steps | number | model default | Diffusion steps (1-50) | | seed | number | — | Fixed seed for reproducibility | | negative_prompt | string | — | Elements to exclude | | cfg_scale | number | model default | Prompt adherence (1-10) | | response_format | url or base64 | url | Output format | | image_url | string | — | Source image for Kontext editing | | image_format | png or jpeg | model default | Output file format | | save_path | string | — | Local path to save output | | disable_safety_checker | boolean | false | Disable NSFW filter |

Use aspect_ratio OR width/height — not both.

Examples

Fast draft:

{
  "prompt": "a cyberpunk cityscape at sunset, neon lights, rain",
  "model": "black-forest-labs/FLUX.1-schnell"
}

Production quality:

{
  "prompt": "portrait of a samurai warrior, dramatic lighting, photorealistic",
  "model": "black-forest-labs/FLUX.1.1-pro",
  "aspect_ratio": "3:4"
}

Edit an existing image:

{
  "prompt": "change the background to a snowy forest, keep the subject",
  "model": "black-forest-labs/FLUX.1-kontext-pro",
  "image_url": "https://example.com/photo.jpg"
}

Save to disk:

{
  "prompt": "game sprite of a blue slime, pixel art",
  "aspect_ratio": "1:1",
  "image_format": "png",
  "save_path": "./assets/slime.png"
}

MCP Client Configs

Claude Desktop

{
  "mcpServers": {
    "together-image-gen": {
      "command": "npx",
      "args": ["-y", "together-mcp-ai@latest"],
      "env": { "TOGETHER_API_KEY": "<YOUR_API_KEY>" }
    }
  }
}

Kiro (~/.kiro/settings/mcp.json)

{
  "mcpServers": {
    "together-image-gen": {
      "command": "npx",
      "args": ["-y", "together-mcp-ai@latest"],
      "env": { "TOGETHER_API_KEY": "<YOUR_API_KEY>" },
      "autoApprove": ["generate_image", "list_models"]
    }
  }
}

Cursor / VS Code

{
  "mcpServers": {
    "together-image-gen": {
      "command": "npx",
      "args": ["-y", "together-mcp-ai@latest"],
      "env": { "TOGETHER_API_KEY": "<YOUR_API_KEY>" }
    }
  }
}

Development

git clone https://github.com/kvginnovate/together-mcp-cg.git
cd together-mcp-cg
npm install
npm run build
npm start

License

MIT