AI Images, Video & More
By spending Cells you can generate images, videos, create talking avatars and remix characters.

None of the models will generate explicit content and prompts which ask for explicit imagery will be rejected.

Float
Float is a versatile image model for anime and realism. You can use text prompts to create images.
Aspect Ratios:
1:1
9:16 or 16:9
3:4 or 4:3
Costs:
Standard costs 4 cells.
Ultra costs 6 cells.
Click the
settings option to select a safety filter rating, with low being more restrictive than high.
GLITCH
GLITCH is great for stylish, creative images.
Aspect Ratios:
1:1
9:16 or 16:9
3:4 or 4:3
2:3 or 3:2
21:9 or 9:21
Costs:
3 Cells
Click the
settings option to select between 1k or 2k resolution and enable or disable enhance prompt which will expand the prompt for better results.
Flux
Flux is a fast, quality model optimized for prompt adherence.
Aspect Ratios:
1:1
4:3 or 3:4
2:3 or 3:2
16:9 or 9:16
21:9 or 9:21
4:5 or 5:4
Costs:
3 Cells on Dev
4 Cells on Pro
6 Cells on Ultra
Click the
settings option to use seeds or enable or disable Raw Mode which makes images more natural looking.
Nano Banana
Nano Banana is great for creative image transformation and editing.
Include up to three reference images and pair with a text prompt to create a new image.
Costs:
4 Cells.
Ideogram
Ideogram is excellent for text and creative designs.
Include up to two reference images and pair with a prompt!
Aspect Ratios:
1:1
1:3 or 3:1
1:2 or 2:1
9:16 or 16:9
10:16 or 16:10
2:3 or 3:2
3:4 or 4:3
4:5 or 5:4
Costs:
3 Cells on Turbo
6 Cells on Balanced
9 Cells on Quality
Click the
settings option to use seeds, style types, and magic prompt options.
Style Types include None, Auto, General, Realistic and Design which controls the aesthetic of the resulting image.
Magic Prompts include Auto, Off, or On which will optimize the prompt to maximize quality and variety.
Seedream
Seedream is a model which creates rich, atmospheric, dreamlike visuals and high-quality, photorealistic images.
Aspect Ratios:
1:1
9:16 or 16:9
3:4 or 4:3
2:3 or 3:2
21:9 or 9:21
Costs:
3 Cells.
Clicking the
settings options will allow you to adjust the guidance scale. Lower values produce more realistic images.
GPT Image
GPT Image is an OpenAI-powered image model.
Include up to three reference images and combine with a text prompt to create new images.
Aspect Ratios:
1:1
2:3 or 3:2
Costs:
1 Cell.
Clicking the
settings option will allow you to adjust the input fidelity between high or low, and change the background between auto, transparent or opaque.
Flux Kontext
Flux Kontext is a frontier model designed for image editing.
Use a single reference image and a text prompt to edit an image.
Aspect Ratios:
1:1
16:9 or 9:16
21:9 or 9:21
2:3 or 3:2
4:5 or 5:4
Or match the aspect ratio of the provided reference image.
Costs:
3 Cells on Base.
4 Cells on High.
8 Cells on Ultra.
Click the
settings option to use seeds or toggle prompt upsampling, which will automatically modify the prompt for more creative generation.

Videos will bring you to the a chat UI where you'll be able to select between two different video generating models.
VEO
You can include one reference image.
Video Resolutions:
720p
1080p
Costs:
320 Cells on Fast.
600 Cells on Standard.
Click the
settings option to include a negative prompt and use a seed.
Seedance
You can include one reference image.
Aspect Ratios:
1:1
4:3 or 3:4
16:9 or 9:16
21:9 or 9:21
Video Resolutions:
480p
720p
1080p
Length:
5 seconds.
10 seconds.
Costs for Lite:
480p: 10-20 Cells for a 5 or 10 second video.
720p: 20-40 Cells for a 5 or 10 second video.
1080p: 35-70 Cells for a 5 or 10 second video.
Costs for Pro:
480p: 15-30 Cells for a 5 or 10 second video.
720p: 30-60 Cells for a 5 or 10 second video.
1080p: 75-150 Cells for a 5 or 10 second video.
Talking Avatar will bring you to a chat interface where you'll be able to combine images with voice clips to create short, talking avatar clips.
Talking Avatars
Combine an image with an audio file to create a talking avatar. The Cell cost varies by audio length.
You can generate an audio clip directly from the menu, using any of the voices available on Xoul.AI or any custom voice assets you've created.

Remixing is an option to take 1-3 Xouls and remix their icon images together into a new image.
Last updated
