POST /v1/images/generations

Generate an image from a text prompt using an image-capable model, with the standard OpenAI images request shape plus a few Openference extensions for diffusion controls.

curl https://api.openference.com/v1/images/generations \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "sdxl-lightning",
    "prompt": "A minimal flat vector icon of a paper airplane, blue on white",
    "size": "1024x1024"
  }'

Response:

{
  "created": 1783970000,
  "model": "SDXL Lightning",
  "data": [
    { "url": "https://api.openference.com/assets/gen/9b6d2c1e-....png" }
  ],
  "usage": { "input_tokens": 13, "output_tokens": 1000, "total_tokens": 1013 }
}

Models

All models below cost 25 plan requests per successful image. See Image generation models for per-model parameters and examples.

Model id	Task	Notes
`sdxl-lightning`	Text-to-image	Fast SDXL-Lightning (default 4 steps).
`sdxl-base`	Text-to-image	SDXL Base 1.0 (default 20 steps).
`sensenova-u1-fast`	Text-to-image	SenseNova U1 Fast (when enabled).

These models appear on the models page and in GET /v1/models.

Common parameters

Field	Type	Notes
`model`	string	An image-generation model (see table above).
`prompt`	string	Required. Text description of the desired output.
`n`	integer	Only `1` is supported.
`size`	string	Optional. Stable Diffusion models: `WIDTHxHEIGHT` (256–2048 px, multiples of 8; default `1024x1024`).
`response_format`	string	`url` (default) or `b64_json`.

Diffusion extensions (Stable Diffusion / Cloudflare image models)

These fields are supported in addition to the OpenAI-shaped parameters above:

Field	Type	Notes
`negative_prompt`	string	Elements to avoid in the output.
`num_steps`	integer	Diffusion steps (alias: `steps`). Max depends on model — see model guide.
`guidance`	number	Prompt adherence (alias: `guidance_scale`). Default 7.5.
`seed`	integer	Optional seed for reproducibility.

Notes

Generated image URLs are hosted by Openference and remain valid — you can hot-link or download them.
Image models cannot be used on the chat, responses, or embeddings endpoints (you’ll get HTTP 400 model_not_supported_on_endpoint), and chat models cannot be used here.
Image requests consume your plan’s request quota using each model’s multiplier (currently 25 plan requests per image for the CF models above).

Getting Started

API Reference

Client Integrations

Models

Billing, Quotas & Usage

API Keys & Security

Reliability & Architecture

Troubleshooting

Reference

FAQ

POST /v1/images/generations

POST /v1/images/generations

Models

Common parameters

Diffusion extensions (Stable Diffusion / Cloudflare image models)

Notes

​POST /v1/images/generations

​Models

​Common parameters

​Diffusion extensions (Stable Diffusion / Cloudflare image models)

​Notes

POST /v1/images/generations

Models

Common parameters

Diffusion extensions (Stable Diffusion / Cloudflare image models)

Notes