TurboClip Canvas · Desktop

One starting point.
A world of possibilities.

Connect images, video, and music in one visual workspace. Build on your best ideas, explore whole sets of variations, or ask your Canvas Agent to plan and build the workflow with you.

Manual building and wiring are free · AI work uses credits
Canvas workflow from a reference image to a board of product variations
01 · Follow the idea

Every result can be
the next beginning.

Start with a reference image. Connect it to a new design, put that design in a lifestyle scene, then explore a whole board of variations. Each result becomes a reference you can carry into the next step.

  • See how an idea develops, with every connection in view.
  • Refine one step or branch into a new direction.
  • Keep the results you love and build on them.
A reference can connects to a new grapefruit design, a lifestyle product shot, and a nine-image variations board.
A reference → a flavor variation → a lifestyle shot → a variations board.
02 · Build the whole set

One product.
Many ways to show it.

A campaign needs more than one good image. Develop a family of product designs, then use those results together as references for new scenes and compositions.

Here, one HALO can becomes three flavor designs. Those designs feed into lifestyle shots, keeping the creative work connected to the same starting point.

  • Combine multiple image references in the next generation.
  • Organize each stage into a named group with its own brief.
  • Explore concepts for launches, social posts, and campaign tests.
One uploaded HALO can branches into three flavour designs, which feed multiple references into lifestyle product shots.
The purple connections show which generated designs guide each new image.
03 · Explore, then refine

Find a direction.
See where it takes you.

Turn one generated image into a board of possibilities. Choose a direction, change the prompt, and use the result as the reference for a more focused next step.

04 · Your Canvas Agent

Talk through the idea.
Build it together.

Ask a question, explore a creative direction, or hand your agent a bigger brief. It can plan the work, create and connect nodes, and build on the images already on your canvas.

“These look great. Now make some social media campaign graphics to match.”

Your conversation stays connected to the work. Review the proposed plan and estimated credits, then see the board take shape. Image stages build on earlier results; video nodes are prepared for you to generate when you’re ready.

The Canvas Agent builds on approved HALO flavour designs with a plan for launch posts, feed and story graphics, and product callouts.
A follow-up request turns approved designs into launch posts, feed graphics, and product callouts.
05 · Beyond the campaign

Same character.
A new adventure.

Bring a character, a piece of artwork, or a playful idea. Explore new looks and settings while using your original image as the reference.

This dog’s photo becomes a set of sunglasses looks, then beach scenes, poolside portraits, and action shots. The same connected workflow works for visual stories, concepts, and creative experiments.

An uploaded dog photograph branches into sunglasses variations and connected beach, poolside, and action scenes.
One familiar subject, connected to a whole collection of new scenes.
06 · Keep the flow going

Drag a connection.
Choose what comes next.

Pull a wire from an image and Canvas shows the nodes that can use it. Create another image, generate a video, explore variations, or collect it in a List.

You can also wire a List into a generator to run once per item. Build a set of prompts, collect the results, and connect the next stage.

Dragging from a dog image opens compatible outputs: Image, Video, Variations, and List.
The connection menu shows compatible next steps for the image you’re working with.
The node palette

A few building blocks.
Your way of creating.

Start with any of these ten node types, then connect them into your own workflow.

  • ImageGenerate an image from a prompt and references.
  • VideoCreate clips from prompts, images, or supported video inputs.
  • VariationsExplore a grid of creative takes from one image.
  • MusicGenerate a soundtrack from a description.
  • UploadBring an image from your computer.
  • AssetsReuse media from your library.
  • TextWrite prompts and reusable instructions.
  • AssistantDevelop ideas or turn a request into text or a list.
  • ListCollect prompts or media and run batches.
  • Sticky NoteKeep the brief and notes beside the work.
The searchable Add node menu, showing image, video, variations, music, and upload options with more below.
Search the Add node menu or browse its Generate, Sources, Text, and Utility sections.
Under the hood

Powered by the best models in AI.

Every generator node reaches the whole bench. Wire an image, clip, or music node to any frontier model, hand the direction to a top text model, and swap any of them the week a better one ships — the graph you built stays exactly the same.

Image nodes
GPT Image 2.5
Nano Banana (3 Pro)
Any frontier image model, up to 4K
Video nodes
Seedance 2.5
MiniMax H3 · Gemini Omni Flash · Seedance 2.0 · Kling V3
AI clips with native audio — edit, reference, extend
Music nodes
Lyria 3 Pro
A score generated straight onto the canvas
Direction
GPT-6 Astra
Opus 5.5 · Gemini 3.1 Pro
The agent + assistant nodes that write every brief
New models added as they launch — pick a different one on any node.
Your next idea starts here

Bring the reference.
Build the possibilities.

Start with one image, connect a workflow, or tell your agent what you have in mind.

Desktop workspace · Manual building is free · AI work uses credits