AI Inspiration WorldWhat it is
OpenUI is an open-source UI generator from W&B (GitHub organization wandb). Describe an interface in words and it renders it live; ask for changes, then convert the HTML to React, Svelte, Web Components and more. The README says W&B also uses it to test and prototype its next generation of tools built on large language models. What sets it apart is the wide model choice: OpenAI, Groq, Gemini, Anthropic, Cohere and Mistral all work, and Ollama runs local models.
Highlights:
- There's a live demo to try first
- LiteLLM connects it to almost any model service
- any OpenAI-compatible endpoint works
- the authors themselves call it an open-source v0 that's not as polished.
How to use:
- First install Docker and set the API keys you need as environment variables.
- Then start it with the docker run command in the README.
- Finally open http://localhost:7878 and start describing a UI.
Who it's for:
Frontend developers and designers who want to try out UI prototypes quickly from a text description.
Before you start:
- Docker is the preferred way to run it; you can also install from source with git and uv.
- Cloud models need their API keys, and your descriptions go to the provider you pick.
- Ollama runs models locally without a cloud call, and the authors warn it will be very slow without a GPU.
- Cloud models charge by usage.
- Apache 2.0 license.
Based on the project's public documentation. Not tested by AI Inspiration World.