Generate images, video, 3D and SVG

Make new images from a prompt, edit an image with a prompt, and create product videos, 3D models and SVG graphics.

The tools in the Create group make something new: an image from a prompt, a prompt-driven edit of a photo, a short video, a 3D model from a product photo, or an SVG graphic. Use them for campaign images, social clips and 3D viewers.

Who can use this:

Tool Permission
Generate image, Transform image, Video generation, SVG generation Use generation tools (image gen, video gen)
Image to 3D Use advanced tools (3D, virtual try-on, lifestyle)

Every plan includes these tools. The run button always shows the exact price of the next run. See AI tools for versions and saving.

Generate image

Generate image makes new images from a text prompt.

  1. Open AI Tools > Generate image.
  2. Describe the image in the prompt box. Two or three words are enough to start.
  3. Optional: pick a model, an aspect ratio and how many images to make. Some models also offer Resolution or Quality.
  4. Click Generate.

Each image is charged. Nothing is saved until you click Keep this image or Keep all. To make many images from a list of prompts, use Batch generate. See Batches.

In the model list Model Credits per image
Fastest drafts (default). Output is 1024 pixels. Nano Banana 2 Lite 2
Sharp images up to 4K Nano Banana 2 5 at 1K, 10 at 2K, 15 at 4K
Quick concepts Z-Image Turbo 2
Detailed prompts FLUX.3 3 at 1K, up to 48 at 4K
Fine detail Seedream 4.5 2
Text and logos Ideogram 4.5 4 at medium quality, less at low, more at high
Photorealism and text GPT Image 2.5 4 at high quality, square. Other qualities and shapes cost less or more

Transform image

Transform image edits an existing photo from a prompt, for example "put the bottle on a wet rock". The photo stays the source, so the product stays recognisable.

  1. Open AI Tools > Transform image.
  2. Add a photo.
  3. Describe the change in the prompt box. Some models also take Reference images.
  4. Optional: pick a model. Some models offer Strength, and Multiple Angles offers Camera controls for rotation, elevation and zoom.
  5. Click Generate.
In the model list Model Credits
Fast edits (default). Output is 1024 pixels. Nano Banana 2 Lite 2
Edits from several images Nano Banana 2 5
Precise edits with text GPT Image 2.5 13
Detailed edits Seedream Edit 4
Text and logo edits Ideogram 4.5 Edit 4
Change only what you ask Bria Fibo 2.0 4
Quick iterations FLUX.3 Edit 3
Quick drafts Z-Image Turbo 1
Place a product in a scene. No prompt needed. Integrate Product 4
New camera angles. No prompt needed. Multiple Angles 4
Remove shadows and glare. No prompt needed. Remove Lighting 4

Nano Banana 2 Lite, Nano Banana 2, Seedream Edit, Ideogram 4.5 Edit and Bria Fibo 2.0 cost the same at any photo size. The other models cost more for photos above 2 megapixels.

Video generation

Video generation makes a short video from a product photo and a prompt. MiniMax H3 Max and HeyGen Video 1 can also make a video from a prompt alone.

  1. Open AI Tools > Video generation.
  2. Add a product photo, unless the model works from a prompt alone.
  3. Describe the motion, for example "a slow turn on a table".
  4. Pick a model, a duration, a resolution and an aspect ratio. Some models can add sound.
  5. Click Generate video.

Video is priced per second. The default, Veo 3.1 Fast for 8 seconds at 720p without sound, costs 80 credits.

In the model list Model Lengths Credits per second
Realistic motion (default) Veo 3.1 Fast 4, 6 or 8 s at 720p, 1080p or 4K 10, or 15 with sound. At 4K: 30, or 35 with sound.
Cinematic motion with sound Kling 3.0 3 to 15 s 8.4, or 12.6 with sound
Same product in every shot. Up to five reference photos. Veo 3.1 Ref 8 s 20, or 40 with sound
4K video LTX 2.3 6, 8 or 10 s 6 at 1080p, 12 at 1440p, 24 at 4K
Premium motion with sound Kling 3.0 Pro 3 to 15 s 11.2, or 16.8 with sound
Premium cinematic shots Seedance 2 Fast 4 to 15 s at 480p or 720p 10.76 at 480p, 24.19 at 720p
From a prompt or one image MiniMax H3 Max 5, 10 or 15 s 5 at 480P, 8 at 768P, 16 at 1080P
Quick clips with audio HeyGen Video 1 5, 8, 10 or 15 s 2 at 480p, 3 at 768p
Talking videos Kling 2.6 Pro 5 or 10 s at 1080p 7, or 14 with sound

Download the result with Download MP4.

Image to 3D

Image to 3D turns a product photo into a 3D model in GLB format, for 3D viewers and AR.

  1. Open AI Tools > Image to 3D.
  2. Add a product photo. A clean photo on a plain background gives the best model.
  3. Optional: pick a model. With Trellis 2 you can set the Resolution, the Texture size and Clean up mesh.
  4. Click Generate 3D model.

The result opens in a 3D viewer on the stage. Click Download GLB to save the file.

In the model list Model Credits
Best balance (default) Trellis 2 22
Everyday 3D models. Textured models from one photo. Seed3D 21
Realistic materials Hunyuan3D 24. Extra options cost more.
Sharp shapes, HD textures Tripo H3.1 38
Detailed 3D models Meshy 7.1 75
Most surface detail Hi3D v3.0 132

The price is the same at any photo size.

SVG generation

SVG generation makes vector graphics, such as icons and logos. It has two modes under Mode.

Text to vector makes an SVG from a prompt:

  1. Open AI Tools > SVG generation.
  2. Under Mode, pick Text to vector.
  3. Describe the graphic.
  4. Optional: pick 1 to 4 Variations, and set Variety. Low follows your prompt; high tries more ideas.
  5. Click Generate SVG.

Vectorize image turns a PNG, JPG or WebP image into an SVG. The image must be under 5 MB and under 16 megapixels, and each side must be from 256 to 4096 pixels.

Mode In the model list Model Credits
Text to vector Icons and logos from text QuiverAI Arrow 15 for each SVG
Vectorize image Turn an image into SVG Recraft Vectorize 1

Download the result with Download SVG.