Hi, I have an app that might work for you under development. It is called the Fluent Digital Twin, and although it's designed for use in video conference calls, it could be used to make videos. The basic idea is that, using AI, the FDT transcribes your speech into text. It turns out that this removes many, but not all, disfluencies. Then the transcription is passed to an AI engine with a prompt to remove the remaining disfluencies, which it turns out is also pretty easy. Finally, the FDT reconstitutes the fluent transcription back into synthetic speech using a cloned voice - which can be made to resemble your own voice. If you would be interested in testing our app, let me know.