Guides Updated

What is Velvet Avocado Studio?

Velvet Avocado Studio brings your AI media project together on your computer: characters, locations, references, prompts, generated takes, and an editor to start assembling them. You choose the generation providers.

Pip the avocado mascot at a desk beside a screen of organised film assets.

The first prompt is easy. Shot forty is the job.

You have a character you like and a shot worth keeping. Now you need the same person from another angle, in the same costume, with the same kitchen behind them. You also need yesterday’s reference image and the settings that produced the good take.

Velvet Avocado gives that work a project home. You can organise reusable characters, locations, props, and costumes, connect generation providers, and revisit the information attached to an output. Changing the model does not mean starting a new filing system.

From a character to a scene

Imagine a short scene in which Evelyn enters an archive and discovers a room full of alternate versions of herself. Here is a way to develop it:

  1. Start a project and collect the references. Keep Evelyn’s approved appearance with her character assets. Organise the archive references as a location so you can find them again for the reverse angle.
  2. Generate a first shot. Choose a connected provider and a model suited to the job. Use your references where that model supports them, and describe the action, framing, and light.
  3. Inspect the take before making more. Open the output’s Generation panel to review its recorded prompt, inputs, and settings. Compare the result with the references you meant to preserve.
  4. Reuse what worked. Use Reuse in composer as a starting point for another attempt. Change the framing or action deliberately, checking which inputs and settings the next model supports.
  5. Start assembling the scene. Bring your chosen media into the built-in Editor to work on the sequence. You can also continue in an external editor when that suits your production.

Velvet Avocado’s Generation panel beside an Evelyn scene, showing the prompt, reference images, settings, and Reuse in composer control.

The Generation panel keeps the recorded prompt and inputs beside the output. You can reuse them as the starting point for another attempt.

That record is useful when an attempt misses, too. You can see what you tried instead of rebuilding it from memory. It does not guarantee that another model will reproduce the same face or composition.

Choose providers within the project

Connect supported services for image, video, and audio generation, or run compatible ComfyUI workflows. Different models support different reference inputs and controls; switching providers still calls for checking those requirements.

The providers page describes the available connections. If ComfyUI is already part of your process, the project organisation guide explains how to keep the workflow and the film organised together.

What local-first means

Your project and saved media live on your computer. You can organise that material locally.

When you choose a cloud provider, the prompts and any media inputs required for that job are sent to that provider for processing. Running a local workflow depends on your ComfyUI setup, including any nodes that call external services. Local project storage does not mean every generation runs offline.

Getting started

Velvet Avocado is free to download and use. Connected providers set their own generation costs; local workflows use your hardware.

The app is available as a preview release for Apple Silicon Macs and Windows. Start with a small project: one character, a location, and a few shots. That is enough to try reference reuse, generation history, and the Editor before moving a larger production across.

Velvet Avocado is free to download and use.