Most voice AI tools lock you into a single provider. You get one LLM, one voice engine, one way to interact. When that provider changes pricing, deprecates a model, or shuts down, you start over.

OpenVoiceUI is an open-source platform that separates the interface from the intelligence. Plug in any LLM -- OpenAI, Anthropic, Groq, Z.AI, or your own local model. Choose any TTS engine, including a bundled free local option. Swap components without rewriting your app.

But it goes beyond voice chat. OpenVoiceUI includes a live web canvas where the AI generates and displays full HTML pages during conversation. Need a dashboard? A report? An interactive tool? The AI builds it in real-time and renders it in the browser. These pages live in a desktop-style interface with windows, folders, right-click menus, and wallpaper customization.

Additional capabilities include AI image generation with FLUX.1 and Stable Diffusion, AI music generation via Suno integration, voice cloning, a built-in music player, and switchable agent profiles that let you run multiple AI personas from one install.

Deployment is Docker-based. Run `npx openvoiceui setup` to scaffold a project, configure your API keys, and launch with Docker Compose. It runs on any VPS or local machine.

OpenVoiceUI is MIT-licensed. The entire stack is open -- no telemetry, no cloud dependency, no usage-based pricing from the platform layer. You bring your own API keys and control your own costs.

Built for developers and tinkerers who want to own their voice AI stack rather than rent it.

GitHub: https://github.com/MCERQUA/OpenVoiceUI