Applio is a powerful, open-source voice conversion platform built on Retrieval-Based Voice Conversion (RVC). Whether you’re converting a single audio file, training a custom voice model, or building a real-time pipeline, Applio provides a full-featured Gradio web UI and a scriptable Python CLI to match your workflow.Documentation Index
Fetch the complete documentation index at: https://mintlify.com/IAHispano/Applio/llms.txt
Use this file to discover all available pages before exploring further.
Installation
Install Applio on Windows, Linux, or macOS using the automated setup scripts or Docker.
Quickstart
Run your first voice conversion in minutes using the web UI or CLI.
Inference
Convert audio files using pre-trained RVC models with full control over pitch, F0 method, and effects.
Model Training
Train custom voice models from your own audio dataset with the four-stage pipeline.
Text-to-Speech
Synthesize speech with Edge TTS and apply voice conversion in a single pipeline.
CLI Reference
Run inference, training, TTS, and utilities headlessly via the core.py command-line interface.
What is Applio?
Applio wraps the RVC voice conversion architecture in an accessible, production-ready package. It provides:- Web UI — A tabbed Gradio interface for inference, training, TTS, real-time conversion, voice blending, downloads, and settings.
- CLI — A full-featured command-line interface (
core.py) for scripting and automation. - Python API — Importable functions for embedding Applio into your own Python applications.
- Plugin system — Extend functionality by dropping
.zipplugin archives into the UI.
Install Applio
Run
run-install.bat (Windows) or run-install.sh (Linux/macOS) to set up the virtual environment and all dependencies automatically.Launch the web interface
Run
run-applio.bat or run-applio.sh. Applio opens in your browser at http://127.0.0.1:6969.Download a model
Use the Download tab to fetch a pre-trained
.pth model and its .index file from a URL or Hugging Face.Key Features
Multiple F0 Algorithms
Choose from
rmvpe, crepe, crepe-tiny, fcpe, or hybrid combinations for pitch extraction.Post-Processing Effects
Apply reverb, chorus, distortion, compressor, delay, bitcrush, limiter, gain, and pitch shift after conversion.
Real-Time Conversion
Convert microphone input to a target voice in real time with sounddevice callbacks.
Voice Blender
Fuse two trained models at a configurable ratio to create entirely new voice characteristics.
Overtraining Detector
Automatically halt training when validation loss stops improving to prevent overfitting.
Docker Support
Deploy Applio in a container with the included Dockerfile and docker-compose configuration.