Applio

About the tool

Applio is a free, open-source (MIT) voice conversion suite built on RVC models and maintained by the IAHispano community. It takes any recording, such as a song, speech or microphone input, and re-sings or re-speaks it in another voice using a voice model. Typical uses are AI song covers in another singer's voice, voiceovers in a different voice and live voice changing for streams and calls. Voice models come from the community or from your own training; Applio itself no longer links to model collections and only offers a Google search for them. The workspace covers audio conversion with pitch, cleanup, upscaling and batch processing, real-time voice changing with low latency, training a voice model from your own dataset, a Voice Blender that mixes several models into a new voice, text to speech run through a voice model, an audio analyzer for datasets and an audio separator that splits vocals from instruments. It runs locally on Windows, macOS and Linux as a desktop app or a Python and Gradio app, with no account and no telemetry, or in the cloud through Google Colab, Hugging Face and Pinokio. Applio is completely free, with no paid plan; the project takes donations on Ko-fi. The MIT license allows commercial use, and the Terms of Use require you to own the audio you process or have permission from its owner. The maintainers say the project is now stable and gets mainly security patches, dependency updates and occasional new features. Support runs through Discord, GitHub issues and support@applio.org.