AICoverGen is an open-source WebUI that creates AI song covers: give it a YouTube video or an audio file and a trained RVC v2 voice model, and it produces a cover of the song sung in that voice. It has about 1,450 stars on GitHub and a pipeline that runs locally or in Google Colab.
Making an AI cover by hand means chaining several tools — download the audio, separate the vocals, convert the voice, remix. AICoverGen packages that whole workflow into one autonomous pipeline with a browser interface. It pulls audio from a YouTube link or a local file, separates the vocal track using UVR-style separation, converts the singing with any RVC v2 trained voice model, and mixes the result back into a finished cover. The project is popular both with hobbyists who want to hear favorite characters sing favorite songs and with developers who want to add a singing feature to an AI assistant, chatbot, or VTuber setup. Setup guides and showcase videos are linked from the repository, and the WebUI can be run on a local machine or through Colab if you do not have a suitable GPU. The repository was last updated in early 2025.
AICoverGen is completely free and open source under the MIT license. There are no plans or fees from the project; your costs are limited to your own hardware, or free Colab resources if you run it there.
Creators, hobbyists, and developers who want to make AI voice covers without stitching together separation, conversion, and mixing tools themselves — especially RVC users who already have trained voice models ready to sing.