What is it?
What it is
Persona is a cross-platform desktop character that provides visual presence during voice conversations, allowing users to customize characters and integrate with voice applications.
Why it exists
To enhance voice interactions by adding expressive visual identity to desktop voice experiences, enabling users to personalize characters and connect with voice apps via MCP.
Who should use it
Developers building desktop voice experiences, users wanting a visual avatar for local AI assistants, and developers integrating VRM/VRMA models into desktop applications.
Who should avoid it
Users looking for a full AI stack (LLM/TTS), developers needing microphone capture (it only captures system audio output), and those without hardware-accelerated graphics.
How it works
A quick walkthrough in plain English
How persona works
Step 1 of 3
You interact with it
Open persona, send a request, or connect it to your stack.
Features
Advantages
- Expressive visual identity enhances voice interactions
- Supports multiple operating systems with specific capture methods
- Easy customization through local model imports
- Seamless integration with Codex for animation control
- No microphone access or network audio transmission for privacy
- Native packaging simplifies distribution across platforms
Disadvantages
- Requires specific software (PipeWire on Linux, Windows 10 build 20348+)
- Initial empty catalog needs manual model imports
- Hardware-accelerated graphics required for rendering
- Packaged media is immutable without user overrides
- macOS requires System Audio Recording permission
- No built-in language models or speech generation
Installation
native
npm install npm run demo
FAQ
What operating systems does Persona support and what are the requirements?
Persona is cross-platform. Linux requires PipeWire and `pw-dump`/`pw-record` on PATH. Windows requires Windows 10 build 20348 or newer for WASAPI loopback. macOS requires version 14.2+ and System Audio Recording permission.
Does Persona capture my microphone or process speech?
No. Persona does not capture the microphone, save audio, produce speech, transcribe content, or send audio over the network. It only listens to the playback process of a supported application to provide visual presence.
How do I use my own 3D models and animations?
You can import your own `.vrm` models and `.vrma` animation files via the Settings menu. Once imported, you can manage them in your local library and assign them to action slots like 'Idle' or 'Speaking'.
How can I connect Persona to an AI agent like Codex?
You can register Persona's local MCP server by running: `codex mcp add persona --url http://127.0.0.1:47831/mcp`. This allows an agent to trigger animations and check the status of the character.
What are the hardware requirements for running Persona locally?
To run Persona from source, you need Node.js 24 or newer, npm, and a desktop session with hardware-accelerated graphics.
Featured in Videos
YouTube tutorials and walkthroughs for persona
Alternatives
Similar projects ranked by category, topics, and text overlap.
