AI

AI Assistant Beta

An AI assistant that builds and edits scenes from a prompt, and generates images on the canvas. Available now, in beta.

Beta

AI Assistant is in beta. It works today, and it's evolving quickly — expect frequent improvements.

Open the AI panel, describe what you want, and the assistant creates or edits objects directly in your project. It understands the full SSHOW design API, so it can build shapes, lay out text, apply styles, and even add scenes and motion so your story plays as you advance.

The AI Assistant panel (Beta): Image, ChatGPT, and Gemini tabs, a model dropdown, an API-key field, a prompt box, and a Go button.
The AI Assistant — three tabs: Image for generation, plus ChatGPT and Gemini for building scenes. Bring your own API key and describe what you want in plain language.

What it can do

Generate scenes Describe a layout in plain language and the assistant builds it on the canvas.
Edit by prompt Ask for changes to what's already there — recolor, rearrange, restyle, add to a scene, or set up reusable {{bound}} variables.
Animate & sequence Ask for transitions and multi-scene sequences — the assistant adds scenes and motion so your story plays as you advance.
Generate images Generate an image from a prompt, or edit one already on the canvas, and drop the result straight onto the active scene.

Tip On the Image tab, pick a model — Nano Banana 2 (fast, the default), Nano Banana Pro (higher quality), or GPT Image 2 — and generate from a prompt. With a Gemini model you can turn on "Use selected image as input" to edit an image already on the canvas instead of starting fresh; OpenAI models add size and quality options.

Getting started

The assistant uses your own API key, so you stay in control of usage and cost.

  1. 1

    Add your API key

    Pick a provider in the AI panel and paste your key. It's stored locally in your browser, never on our servers.

  2. 2

    Describe what you want

    Type a prompt — be as specific as you like. With Gemini, adding 'step by step' nudges it to think more carefully. Press Shift+Enter or Cmd+Enter to run, or click Go.

  3. 3

    Review the result

    The assistant applies its work to your scene. Not quite right? Refine the prompt and try again, or undo.

Providers & models

Choose the provider that fits your needs. You can switch any time — the assistant remembers each provider's key, last model, and the tab you were on, all in your browser.

OpenAI ChatGPT models for scene generation, plus image generation.
Google Gemini Gemini models with adjustable thinking depth, plus fast, low-cost image generation and image-to-image editing.

Privacy & your keys

Note Your API key lives only in your browser's local storage. Prompts and project context go directly to the provider you chose under your own account — we don't proxy or store them.

Read the AI Usage Notice