Text-based video editing: cut speech like a document

Transcription runs locally with whisper.cpp, with models you can download in Settings, so nothing is uploaded. If you prefer, pick OpenAI for transcription instead in Settings → AI & models.

Select words and delete them to cut them from every track, keeping picture and sound in sync. Filler words go in one click, and long pauses can be removed across the timeline.

Find a moment by what is said, even when you do not remember the exact words, generate chapters, and export captions as SRT or VTT. Search by meaning and chapters use Ollama or LM Studio on your Mac, or OpenAI.

An agent in Cue's Agent panel or any MCP client can do the same, for example: "Transcribe the interview, cut the filler words and add captions." Each cut is a normal, undoable edit on the timeline.