Visuali Review: An AI Image Editor on an Infinite Canvas, With Tokens to Watch

Visuali is an AI image generator and editor that works on an infinite canvas instead of a chat box. You generate with a prompt, then move, mask, extend and redraw the result as layers. It is the way a designer works in Photoshop or Figma. It is built by Visuali AS, a small company in Oslo, Norway, and has been around since early 2023. It runs in any browser and installs as an app on phone or desktop.

What does Visuali do?

The canvas is the product. Every generation lands as a layer you can drag, resize, rotate, hide or delete. Generate makes images from text with a choice of models. Edit changes only what sits inside a frame you position over the image. Fill and Extend erases and replaces elements or pushes the picture past its edges. Multiple Inputs feeds up to five reference images into one edit. You can combine a product shot, a background and a style sample. There are brushes, an eraser, a colour pipette, text with several fonts and shapes.

The models are third party, and the list is current. It includes GPT Image 2 and 1.5, Nano Banana and Nano Banana 2, Gemini 3 and Seedream 4.0. Flux 2, Flux 2 Pro, Flux Schnell and Dev, Stable Diffusion XL and Z-Image are there too. One subscription covers all of them.

An AI agent, in beta, sits beside the tools. You tell it "make a summer promo from these" or "remove the background". It then works on the canvas: generating, cutting out subjects, adding text and laying out the result as editable layers. It can also produce assets in bulk, such as one product in ten variations or every size for a launch.

Projects save to the cloud. You can export layers, download images and reopen work on another device. Visuali installs from the browser as a full-screen app on iPhone, Android and desktop. There is nothing to download from a store.

Which features stand out?

• Infinite canvas with layers — every generation and edit stays editable, which no chat-style generator offers.
• Frame-based editing — inpainting and outpainting by positioning a frame, not by writing masks in a prompt.
• Multiple reference inputs — up to five images as context for one edit, good for compositing and brand consistency.
• Current model roster — GPT Image 2, Nano Banana 2, Seedream 4.0 and Flux 2 Pro behind one bill.
• AI agent in beta — a chat that works on the canvas and can produce whole campaigns in one brief.
• Installs as an app — full-screen on phone, tablet and desktop from the browser.

How much does Visuali cost?

Three self-serve plans plus a custom tier, all on monthly tokens. Prices below are monthly, with the annual toggle marked "Save 50%".