Model
Spotlight search
Find any model or tool and jump to its page.
Pay-per-use workspace for AI image, video, vector, 3D, and audio generation.
AI models by provider
vsb CLI · plays nice with Claude Code
Make anything from your terminal
Install the vsb CLI and every model here is one command away. Better yet, let Claude Code drive: tell it what you want and it runs the commands for you. Everything lands in your library.
curl -fsSL visualsandbox.com/install.sh | sh && vsb setupClaude Code
>make a neon jellyfish drifting through a dark ocean
●Bash(vsb run image/nano-banana --prompt "neon jellyfish…")
✓Done in 6s
●Bash(vsb run video/p-video --image out.png --prompt "slow drift…")
✓Done in 12s · saved to your canvas
$
How it works
- Step 1
Pick a model
Choose from the latest models across image, video, vector, and 3D. New ones added regularly.
- Step 2
Write a prompt
Describe what you want. Adjust settings if you need finer control. Attach references to guide the result.
- Step 3
Pay only for what you make
No subscription. No credits to expire. Card on file, charged per generation. You see the price before you make it.
Frequently asked
Visual Sandbox is a pay-per-use platform for creating with AI. Image, video, vector, and 3D generation in one workspace, with the latest models from Google, OpenAI, ByteDance, Tencent, ElevenLabs, and others.