Unknown · 2× · vu le 18 août 2026
26
Momentum
Persona is an open-source desktop application (Electron) that displays an animated 3D VRM character on screen and drives it in real time via lip-sync and body animation whenever a voice AI (e.g. ChatGPT or Codex) produces audio output. The app detects speech purely through system audio capture (RMS amplitude), does not run language models itself, and connects AI agents via a local MCP server that exposes animation controls. The source code is released under the MIT license; bundled character models and animations are subject to separate licensing terms.
Historique du momentum
20.05.18.08.
Fonctionnalités
| Autonomy Level | Reactive: automatically triggers animations on detected speech output; does not run language models itself |
| Channels | Standalone desktop app running alongside voice AI tools like ChatGPT or Codex |
| License | MIT license (app source code); character models/animations licensed separately |
| Modalities | Audio detection (system output) and 3D visualization (VRM avatar with lip-sync and body animation) |
| Platform | Desktop: Windows (NSIS), macOS (DMG/ZIP, ARM64+x64), Linux (AppImage, DEB) |
| Price | Free, open source |
| Release Date | Initial release approx. late July 2026 (680 stars in 3 days, as of 3 weeks ago) |
| Tool/MCP Integrations | Own local MCP server (port 47831); compatible with Codex and other MCP clients for animation control |
