resolution

Faster generation for high-quality everyday content.

GPT Image 2.5

Posts, packaging, and UI mockups with sharp text.

What you want to generate or how to edit the input.

One or more reference images (for editing or composing).

Generation quality. Lower quality is faster and cheaper.

Optional: bring your own OpenAI API key to pay OpenAI directly.

gpt image 2

Nano Banana 2.1

Professional-grade images with finer details.

Seedream 5 Pro

Your product, accurately placed across bulk images.

Text prompt to generate or edit an image.

Pass in one or more images. Use up to 14 reference images for style transfer, image fusion, and complex editing.

Automatically match the aspect ratio of your input image.

Output resolution. Note: exact API key name not documented on the public model page.

Use Google Image Search grounding to find web images as visual context for generation. When enabled, web search is also used automatically.

Use Google Web Search grounding to generate images based on real-time information (e.g. weather, sports scores, recent events).

nano banana 2

URL of a reference image to guide generation.

Weight of the reference image; larger values give it stronger influence.

photon

Realistic visuals with accurate text, in any language.

The text prompt to generate an image from.

The seed for the random number generator.

Content moderation strictness: 1 = strictest, 6 = least strict.

System instruction steering the model's persona and output style.

If true, the image is returned inline as a data URI (not stored in request history).

Experimental: limit each prompting round to one generation; overrides in-prompt count instructions.

Allow the model to use live web information when generating. Adds $0.015 per request.

Nano Banana Pro

Sharper images with smarter prompt understanding.

The text prompt used to generate the image.

Number of separate model generations to be run with the prompt.

If set to true returns trusted URLs. Please note these URLs expire in 24 hours.

seedream 5 lite

Accurate brand colors and sharper details in every image.

Recraft v4.1

Create dozens of post and Story layouts in seconds.

Style reference image to guide generation.

Character reference image to guide generation.

photon-flash

Professional images with consistent style.

Recraft v4 normal

The fastest and cheapest way to create images.

Denoising steps. 4 is recommended; fewer = faster, lower quality.

Save quality 0–100 (not relevant for .png outputs).

Approximate megapixels for the generated image.

Faster fp8-quantized inference; non-deterministic even with a seed.

Disable safety checker for generated images.

flux-schnell

Memes, test posts, and background swaps, fast & cheap.

Name of the S3 bucket that holds the uploaded media (key / img_keys / mask_keys reference objects in this bucket).

Text prompt for image generation or editing.

Optional input image for editing. When provided, the model edits this image based on the prompt. S3 object key(s) of the input image(s) in the bucket (array).

Aspect ratio of the generated image. Ignored when editing an image.

Grok Imagine Image

Size of the generated image. Width/height 1920-4096, or total pixels 2560x1440-4096x4096.

Number of separate model generations to run with the prompt.

If >1, enables multi-image generation per generation.

Random seed to control the stochasticity of image generation.

If true, media returned as a data URI and not stored in history.

ByteDance Seedream 4.5 Text to Image

High-volume images with brand consistency.

Recraft V4.1 Utility Pro

Great for accurate face handling and fine textures.

The text prompt to guide image generation.

Check input for safety before processing.

Check output for safety after generation.

Acceleration level (faster = lower quality).

Wan 2.2 A14B Text to Image

Size of the generated image. Total pixels must be between 960x960 and 4096x4096.

Prompt-enhancement mode. Standard: higher quality/slower; Fast: average/faster.

ByteDance Seedream 4.0 Text to Image

Recraft v4 pro

Restyle your images while text stays readable.

https://cdn.myzesty.com/resources/ai_icon/web-aspectratio.png

CFG scale - how closely the model follows the prompt.

Safety tolerance (1 strictest, 6 most permissive).

Whether to enhance the prompt for better results.

FLUX.1 Kontext [pro] Text to Image

If true, media returned as data URI and not stored in history.