← All models
Edits and combines reference images using text instructions, with optional prompt expansion, output-size controls, and reproducible generation through a seed.
images
Details
- Accepts
- text
Pricing
- Price
- 3.13 cr / megapixel
Prices in credits (1 credit = $0.01).
Data schema
Input
| Field | Type | Description |
|---|---|---|
| seed | — | Latent noise seed. Use a new seed when editing a previous output again. |
| model | stringenum: berry, jam | Model variant: berry favors quality, jam favors speed. |
| promptrequired | string | Description of the image to generate. |
| sync_mode | boolean | Return images as data URIs instead of saving them in request history. |
| image_size | — | Output size, rounded up to multiples of 16, at most 4,220,160 total pixels (1680×2512). auto lets the prompt expander pick the aspect ratio at about 1024×1024. |
| image_urls | array | Reference images to edit or combine (1-8). Order sets the <image-N> tags. |
| output_format | stringenum: png, jpeg | Output image format. |
| negative_prompt | string | What to keep out of the image. Used by berry only. |
| reasoning_level | stringenum: low, medium, high | How much effort the harness spends reasoning about the given editing task. Higher reasoning effort means slower generation time. high plans the edit with segmentation, web search and reference photos, keeps unedited areas of a single input image, corrects required text, and bills 5 megapixels more p… |
| enable_safety_checker | boolean | Check output images for unsafe content. Disabling requires account authorization. |
Output
| Field | Type | Description |
|---|---|---|
| seed | integer | Latent noise seed used for generation. |
| images | array | Generated images. |
| prompt | string | Prompt used for generation, after expansion. |
| timings | object | Seconds spent in each generation stage; `inference` is the total GPU sampling time. |
| has_nsfw_concepts | array | Safety result for each image. |