> ## Documentation Index
> Fetch the complete documentation index at: https://docs.ekly.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Images

> Which image model to reach for, and the fields that matter.

Call `GET /v1/models` and filter `type: "image"` for the live list and prices. A few starting points:

| You want | Start with |
| - | - |
| Fast, cheap drafts and variations | `gemini-2.5-flash-image` |
| Highest fidelity, text in the image, transparent backgrounds | a GPT Image 2.5 model (`image_quality`, `output_format`, `image_background`) |
| Edit or restyle an existing image | any model with `reference_images`; add `mask_image` for inpainting |

Fields that apply across models: `prompt`, `negative_prompt`, `aspect_ratio`, `generate_image_count`.
Model-specific ones (quality tiers, resolution, formats) are listed with their allowed values and
a ready-to-send `example_request` in `GET /v1/models/{id}` and on each page under **Image models**;
sending a value the model does not accept is a `422` before any charge.

Prompting advice for specific models lives on ekly.ai, for example the
[Kling prompting guide](https://ekly.ai/guides/kling-prompting).


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.