MorphVox — real-time voice conversion, TTS, and song voice-swap (RVC as a usable web platform)
Site: https://morphvox.net
GitHub: https://github.com/alli959/rvc_real_time
MorphVox is my attempt to take the RVC ecosystem and turn it into something you can actually use as a product: a web app where you can browse voice models, convert voice in real time, generate TTS, and do full song voice swaps — all in one place.
What you can do
- Real-time voice conversion (streaming) with RVC models (
.pth+.index) - Text-to-speech with lots of voices + styles/emotions, and you can run the result through any model
- Song Remix + Voice Swap (new lead singer workflow)
- Model ownership + access control (private models, or share access with specific users)
Two ways to use it
- One model → try everything: go to /dashboard/models, select a model, and test different features/settings with that same voice.
- One feature → try many models: open a specific page (TTS / voice generation / remix) and quickly switch models from the dropdown.
Song Remix / Voice Swap (how it works)
- Search for a song in the Remix search field (searches YouTube) and download the track as an MP3
- In Voice Swap, choose a model and click Transfer
- The output comes back with a new lead singer (the model you selected)
- If the voice doesn’t fit the song, you can raise/lower pitch to make it sit better
Wizart (in development)
Wizart is a guided creator flow:
- Create your own voice through a step-by-step wizard
- Improve existing models by feeding voice clips based on how they perform
- After tweaks, the model’s performance is calculated again so you can track improvements
Run it locally (Docker)
The repo includes a full local setup under infra/compose so you can run the whole stack on your machine (web app + API + voice engine + storage) using Docker Compose.
Local voice chat (Discord-style) — local instance only
This is not available on the production site, but in a local setup you can use the voice-engine as a real-time streaming service and route the converted audio into Discord using a virtual microphone/audio routing tool. (Pick a model → stream mic audio → receive converted output → select it as your Discord input device.)