HiDream-O1-Image API

HiDream.ai's open-weights HiDream-O1-Image (8B): one pixel-native model that generates, edits, and personalizes without a VAE or a separate text encoder.

Endpoint

POST https://api.hedra.com/v3/models/hidream-o1-image

Example request

curl -X POST "https://api.hedra.com/v3/models/hidream-o1-image" \
  -H "Authorization: Key $HEDRA_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"input": {"prompt": "a cinematic shot of a city at dusk"}}'

Parameters

promptstringrequired

Generation prompt. At least 1 character.

num_outputsinteger

Number of outputs generated per job. Only 1 is supported.

enhance_promptboolean

Rewrite the prompt before generation. An LLM expands it into a fuller description and the model receives that text instead of the submitted one; the result's `prompt` reports what ran.

aspect_ratiostringrequired

Output aspect ratio.

imagesarray

Images to edit or blend. 1 to 10 images, each at most 30 MB.

output_formatstring

Output image format.

seedinteger

Seed for reproducible output; omit for a random seed.

guidance_scalenumber

How closely the model follows the prompt. From 0 to 20.

num_inference_stepsinteger

Denoising steps to run. From 1 to 50.

resolutionstring

Output resolution.

qualitystringrequired

Quality level to generate at.