Have recordings with spoken words cut down to what matters. Filler words, long pauses and repeated sentences disappear, and the recording keeps its flow. Nothing is cut until you have confirmed the cutting strategy in plain words.

The Agentic Part

An agent first turns the recording into a transcript, the spoken text with the time of every single word, produced locally with no cloud service. From it comes a compact overview in which the agent marks filler words, long pauses and repeated sentences. It decides what stays: for a repeated sentence the last spoken version counts, and a long pause is shortened to a brief breath. Cuts land in silence and never in the middle of a word. At every cut the agent checks picture and sound track itself, so no click and no clipped word remains. It proposes the cutting strategy to you, and the recording is cut once you have confirmed it.

Main Features

Filler words, pauses and repetitions removed from spoken recordings
When a sentence was spoken several times, the last spoken version stays
Cuts in silence, never in the middle of a word
Long pauses are shortened and kept as a brief breath
Smooth sound transitions at every cut
Every cut checked on picture and sound track before you see the result
Burned-in subtitles, uniform loudness and a colour match per scene are possible on top
The transcript is produced locally, with no cloud service

Use Cases

You have the recording of a screen demonstration with throat clearing, pauses and restarts. The agent cuts them out, and you receive a smooth version.
A sentence was spoken three times until it sat right. The cut keeps the last version.
The video should get subtitles. They come from the transcript and are burned into the picture.