AI generation
Virse generates images and text directly on the canvas. The generation interface appears contextually depending on what's selected.
Generating into an element
Select a single image or text element. A chat bar appears at the bottom of the canvas.
Type a prompt and press Enter or click Generate. The element shows a loading state while generation runs, then updates in place with the result.
You can also drag reference images from elsewhere on the canvas into the chat bar to use them as visual inputs.
Generating a batch
With nothing selected, the anchored chat bar at the bottom lets you generate multiple images at once. Set the batch size, choose a model, and generate — the results are placed on the canvas and automatically positioned to avoid overlap.
Models
Virse supports multiple image generation models. The right model depends on your goal:
Fast iteration
Nano2
High quality output
Nano Pro or Imagen 4 Ultra
Complex instructions or text rendering
GPT Image
Style transfer from a reference
FLUX Kontext
Precise resolution control
FLUX 1.1 Pro Ultra
Use List models in the model selector to see the full current list with supported parameters.
Parameters
Model — which generation model to use
Aspect ratio — controls the output dimensions
Style — apply a trained style from your library
Aesthetic memory — let LAMM steer the output toward your taste profile
Inpaint, outpaint, and upscale
These are targeted editing operations available from the toolbar when an image element is selected:
Inpaint — paint over a region to regenerate just that area
Outpaint — extend the image beyond its current edges
Upscale — increase the resolution of the image
Last updated