FirstCut.video

FirstCut.video is a transcription tool. FirstCut is a browser-based narrative editor that turns interview and talk footage into editable beats, clips and words, with an AI assistant, revisions and export to professional editors. Only audio is uploaded.

FirstCut is a browser-based narrative editor for footage where the story lives in what people say. Rather than opening on a timeline, it reads the spoken material and organizes it into a story made of beats, clips and words. A beat is a narrative point built from consecutive clips, a clip is a span of source footage with its matching transcript text, and the word level lets an editor cut individual lines or phrases while the surrounding story stays visible. The tool is aimed at editors, producers and content teams working with interviews, documentaries, podcasts, courses, lectures and webinars. In practice, footage is connected, the transcript is mapped, and a first narrative appears as a readable outline. From there the editor can drag beats into a new order, rename, add or delete them, attach instructions to a beat or clip, and generate an entirely different story arc from the same material while the original stays intact. An AI assistant accepts plain-language requests such as tightening an intro, removing repetition or lengthening the sequence, and turns them into changes on the sequence. Each meaningful change is saved as a revision, so earlier versions remain reachable, and every revision can be played back directly in the browser. Privacy and setup are handled differently from most AI video tools. Nothing is installed and no GPU is required. The browser extracts only the audio, which is uploaded to produce the transcript, and the video itself stays on the local computer. The AI works from the transcript rather than the picture, and it selects and orders existing moments instead of generating images, video or audio. FirstCut positions itself as a rough-cut stage, not a replacement for a finishing editor. Finished sequences export with source footage and timecodes preserved, in FCPXML, FCP7-XML, CMX3600 EDL and OTIO formats, for Premiere Pro, DaVinci Resolve, Final Cut Pro or Media Composer. It suits speech-led material far better than visually driven footage, and it sits between transcript-based editors and full nonlinear editing suites.