About OpenVoiceUI
Not another chatbot. A voice-first AI platform that builds web pages, manages files, generates images, writes code, and runs autonomous workflows — while you talk to it.
OpenVoiceUI started because every voice assistant felt the same: you talk, it talks back, and nothing actually happens. You could ask a question and get an answer, but you couldn't say "build me a dashboard" and watch it appear. You couldn't say "research my competitors and write a report" and come back to a finished document.
The missing piece was a visual canvas. Voice is great for intent, but most real work produces artifacts — pages, images, charts, files. So we built a system where your AI doesn't just talk, it builds. The canvas is a full desktop environment: file explorer, taskbar, live HTML pages that the AI constructs during conversation. You speak, things appear.
Under the hood, OpenVoiceUI runs on the OpenClaw gateway — a programmable AI backend that handles LLM inference, tool execution, sub-agent orchestration, and session management. This means OpenVoiceUI isn't locked to any single AI provider. Swap models, add tools, change behavior — all through configuration, not code changes.
I'm a developer and entrepreneur in Arizona who believes AI should be a tool you own, not a service you rent. I built OpenVoiceUI because I wanted a voice AI that could actually run my business operations — manage clients, build websites, create marketing content, schedule tasks — without locking me into any single vendor's ecosystem.
The project is open source because I think the best tools are the ones people can inspect, modify, and make their own. If you host it yourself, your data stays on your machine. If you want managed hosting so you don't have to think about infrastructure, that's available too. Either way, the code is yours.
Use OpenAI, Anthropic, Groq, local Ollama models, or anything else. Swap providers without changing your application. Your AI workflow shouldn't depend on one company's API staying affordable.
Your conversations, your data, your server. OpenVoiceUI runs on your own infrastructure with Docker. No cloud dependency, no data leaving your network unless you want it to.
No restrictions on commercial use. Fork it, modify it, build a product on it, sell it. The MIT license means you can do whatever you want with the code.
Issues, PRs, and feature requests are welcome. The project grows based on what real users need, not a product roadmap designed to upsell you on a premium tier.
Everything is on GitHub. Clone it, read it, break it, improve it.