Find the fillers. Keep your voice.

Review before you remove.

Open the filler tool, choose the language and run detection. Review the suggested ranges with the surrounding speech. A filler may be part of someone's voice or meaning; keep it when the sentence needs it.

What runs locally and what connects.

Transcription uses the local speech engine when available. If that engine is missing, this tool can upload compressed audio for cloud transcription. Detection sends the transcript to the CleanCut service. An active Pro license and connectivity are required.

The final video processing stays on your computer. This is a different workflow from MCP's explicitly local transcription, which does not fall back to a cloud upload.

Finish with a listening pass.

Listen for cut words, unnatural breaths and changes in meaning. Adjust the selection and export a separate version so the original stays available.