Editing narration often meant repeating the same mechanical task: find a failed take, cut it, compare the next attempt, and remove the gaps that no longer belonged. The work was simple in principle, but slow in practice—especially for recordings with many corrections and restarts.
I tried existing tools, but none fully matched the workflow I wanted, and several useful options were paid. I wanted a tool I could inspect, adapt, and run with source media kept on my own machine.
That led me to build VoiceCut, an open-source, local-first project for turning retake-heavy narration into a cleaner audio or video edit. It is designed around selecting good takes from the original recording rather than synthesizing replacement speech.
VoiceCut is still evolving. If this workflow sounds useful, you can try it, test it on your own recordings, report an issue, or contribute through the repository.