OpenVoiceUI

The open-source voice AI that actually does work.

Install, open localhost:5001, say "build me a dashboard", and watch it render live.

Watch the demo -- see voice-to-canvas in action

Install

Prerequisite: Docker must be installed and running for all install methods.

Pinokio (one-click)

Download Pinokio if you don't have it, then search "OpenVoiceUI" in the app store and click Install.

npm

npx openvoiceui setup     # interactive wizard — walks you through API keys + builds Docker images
npx openvoiceui start     # starts everything

Docker

git clone https://github.com/MCERQUA/OpenVoiceUI.git
cd OpenVoiceUI
cp .env.example .env        # edit with your API keys
docker compose up

Open localhost:5001 and start talking.

What is OpenVoiceUI?

OpenVoiceUI is a hands-free, AI-controlled computer. You talk — it builds. Live web apps, dashboards, games, full websites — rendered in real time while you watch. No mouse, no keyboard, no typing prompts into a chat box.

It runs on OpenClaw and works with any LLM. The AI agent can build and display apps mid-conversation, switch between projects with a voice command, generate music on the fly, delegate work to parallel sub-agents, and remember everything across sessions. It uses any Claude Code or OpenClaw skill — and the community can build and share more through the plugin system.

Self-hosted. Your hardware, your data. MIT licensed, forever free.

Core Features

• Hands-Free AI Computer — Talk and watch it work. The AI builds apps, switches between projects, runs tasks, and displays results on a live visual canvas — all without touching a mouse or keyboard.
• Live Canvas — AI renders real HTML pages mid-conversation: dashboards, tools, galleries, reports, full web apps. Not text responses — real interactive pages you can use.
• AI Music Generation — Generate songs on the fly with your voice using Suno. Full music player with playlist management built in.
• Custom Animated Interface — Choose from animated face modes (eye-face avatar, reactive halo-smoke orb) or install community-built faces through plugins. Build your own — the face system is fully extensible.
• Sub-Agents — Delegate multiple tasks to parallel AI workers simultaneously and get results back.
• Long-Term Memory — Optional context engine plugin curates knowledge every turn. Persists across sessions in human-readable markdown.
• Desktop OS Interface — Themed desktop environment with window management (Windows XP, macOS, Ubuntu, Win95, Win 3.1).
• Admin Dashboard — Mobile-responsive. Agent profiles, provider config, workspace file browser, plugin management, system health. Everything editable live.
• Self-Hosted — Your hardware, your data. No vendor lock-in, no monthly fees.

And More

• Image generation (FLUX.1, Stable Diffusion 3.5)
• Video creation (Remotion Studio)
• Voice cloning (Qwen3-TTS via fal.ai)
• Cron jobs for scheduled automation
• File explorer with drag-and-drop
• Agent profiles — switch personas, voices, and LLM providers from the admin panel

Install Details

Option 1: Pinokio (one-click)

1. Install Pinokio if you don't have it
2. Search "OpenVoiceUI" in the Pinokio app store
3. Click Install, then Start

Pinokio handles Docker, dependencies, and configuration automatically.

Option 2: npm

Requires Node.js 20+, Python 3.10+, and Docker.

npx openvoiceui setup     # interactive wizard — configures LLM, TTS, API keys, builds Docker images
npx openvoiceui start     # starts OpenClaw gateway + Supertonic TTS + voice UI

The setup wizard walks you through choosing an LLM provider, TTS provider, and entering API keys. Configuration is saved to .env and openclaw-data/.

npx openvoiceui stop      # stop all services
npx openvoiceui status    # check what's running
npx openvoiceui logs      # tail service logs

Option 3: Docker

Requires Docker and Docker Compose.

git clone https://github.com/MCERQUA/OpenVoiceUI.git
cd OpenVoiceUI
cp .env.example .env

Edit .env with your API keys (at minimum: an LLM provider key and optionally a TTS key). Then:

docker compose up -d

This starts three containers:

Open http://localhost:5001 to use the voice interface, or http://localhost:5001/admin for the admin dashboard.

To stop: docker compose down

Option 4: VPS / Production

For running on an Ubuntu server with nginx and systemd:

git clone https://github.com/MCERQUA/OpenVoiceUI.git
cd OpenVoiceUI
cp .env.example .env               # edit with your API keys
sudo bash deploy/setup-sudo.sh     # creates dirs, installs systemd service
bash deploy/setup-nginx.sh         # generates nginx config (edit domain)

See deploy/ for the full production setup including SSL, nginx reverse proxy, and systemd service files.

Configuration

All configuration is in .env. Copy .env.example to .env and fill in your values.

Required:

• An LLM provider API key (OpenAI, Anthropic, Groq, Z.AI, or any OpenClaw-compatible provider)
• CLAWDBOT_AUTH_TOKEN — set during npx openvoiceui setup or in OpenClaw's setup wizard

Optional but recommended:

• GROQ_API_KEY — enables Groq Orpheus TTS (fast, high quality, free tier)
• SUNO_API_KEY — enables AI music generation
• CLERK_PUBLISHABLE_KEY — enables login/auth (for multi-user or public deployments)

See .env.example for all available options with descriptions.

Works With Any Provider

LLM

Text-to-Speech

Speech-to-Text

Admin Dashboard

Access at localhost:5001/admin. Mobile-responsive.

• Profiles — View and activate agent personas
• Agent Editor — Edit name, voice, LLM provider, system prompt, features, and agent workspace files. 4 tabs: Profile, System Prompt, Features, Agent Files
• Plugins — Install and manage face packs, gateways, and extensions
• Canvas Pages — Toggle public/private, lock pages, delete with archive
• Workspace Files — Browse and edit agent workspace. Audio playback, image preview built in.
• Music (Suno) — View all generated songs, play inline, archive tracks
• Provider Config — Select LLM, TTS, STT providers. Saves to active profile.
• Health and Stats — CPU, RAM, disk, gateway status, session reset
• Connector Tests — 12 automated endpoint diagnostics

Use Cases

Small Business — AI receptionist, appointment scheduler, report builder. Talk to your AI and get a live dashboard of today's leads, reviews, and tasks.

Digital Agencies — Deploy custom AI assistants per client. Multi-tenant ready. Each client gets their own voice-powered workspace.

Developers — Fork it, extend it, deploy it anywhere. MIT licensed. Build custom plugins, gateway adapters, and canvas pages on top of a voice-first platform.

How It's Different

Tech Stack

Plugins

OpenVoiceUI has a plugin system for community-built extensions. Plugins can include animated face packs, canvas pages, workflow dashboards, gateway adapters, or any combination.

Build your own. Face packs, canvas pages, workflow dashboards, gateway adapters (template), or STT/TTS adapters (template). See the plugins repo for submission guidelines.

Documentation

• Introduction
• Environment Variables
• Plugin Development
• Contributing
• Website

Contributing

We welcome contributions — especially plugins. Build a face pack, a canvas page, a workflow dashboard, or a full extension and submit it to the plugins repo. See CONTRIBUTING.md for code contribution guidelines and openvoiceui.com for full documentation.

License

MIT

Website  ·  GitHub  ·  npm  ·  Plugins