What is LUVI and what can I generate?
LUVI is one account and one credit balance for 223 AI models — images, video, speech, music and 3D — with a shared library, workflows, and a connector that lets Claude and ChatGPT generate for you.
What can I make?
Four kinds of output, each from several model families:
- Images — text-to-image, image editing and multi-image composition (Seedream, GPT Image, Nano Banana, Grok Imagine, Qwen-Image, FLUX and more).
- Video — text-to-video, image-to-video, reference-to-video, video edit and extend, most of it with sound generated in the same pass (Seedance, Veo, Kling, MiniMax H3, Wan, Gemini Omni, HappyHorse, PixVerse, Vidu and more).
- Audio — text-to-speech, music and speech-to-text.
- 3D — image-to-3D and text-to-3D meshes.
Every model has a public page with its parameters, its starting price in credits and the prompting guide we use ourselves.
How does it fit together?
The Workspace is where you pick a model, write a prompt, attach references and generate. Every result lands in your Library, organised by project and folder, with tags, ratings and a player built for the files these models produce. Workflows chain models together on a canvas — one image feeding a video, one reference reused across shots — with a cost estimate before you press Run.
What does it cost?
You pay in credits. Every model page says what a generation starts from, the Workspace shows the estimate before you confirm, and the final charge is settled against the provider's real price — never above the estimate. New accounts start with 1,000 credits.
Where do I start?
Sign up, open the Workspace, and generate an image with the default model. The next article walks through it step by step.
Frequently asked
Do I need separate accounts for each model provider?
No. One LUVI account and one credit balance cover every model on the platform.
How many models are there?
223 public models across 30 families at the time of writing: 138 video, 63 image, 18 audio and 4 3D models. The live list is at /models.
Can I use LUVI from Claude or ChatGPT?
Yes. LUVI is an MCP connector: add it once and both assistants can list models, estimate a cost, generate, and read your library.