Comfy Cloud Image Models

These image models run through the Comfy Cloud account pool. Clients do not need to know about the pool or workflow: select a model and use its official request and response format.

Unified image generation API

Use POST /v1/images/generations for generation and POST /v1/images/edits for editing. FocalAPI preserves the creative parameters required by each model while returning a unified image result.

curl https://api.focalapi.com/v1/images/generations \
  -H "Authorization: Bearer $FOCALAPI_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "doubao-seedream-5-0-pro-260628",
    "prompt": "A lighthouse by the sea in morning fog, cinematic lighting",
    "size": "1024x1024",
    "n": 1
  }'

gpt-image-2 supports low, medium, and high quality, plus background and single-image mask editing. Supported sizes, reference images, and output counts vary by model; follow the API validation response.

Native Gemini API

Use Gemini's native generateContent endpoint for the following models. To generate an image, generationConfig.responseModalities must contain only IMAGE.

curl 'https://api.focalapi.com/v1beta/models/gemini-3.1-flash-image-preview:generateContent' \
  -H "Authorization: Bearer $FOCALAPI_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "contents": [{"parts": [{"text": "A clean product illustration of a blue ceramic mug"}]}],
    "generationConfig": {
      "responseModalities": ["IMAGE"]
    }
  }'

Images are returned as inlineData in response candidates. Use Gemini-native aspectRatio and imageSize when available; FocalAPI validates them against the selected model.

Billing and retries

Pricing depends on request parameters such as image count, resolution, quality, and reference-image count. FocalAPI reserves quota before submission and settles against successful output count; a failed submission does not retain a charge. Check the Model Marketplace for current prices.