SMF·DESCRIPT
Can a prompt replace Descript?
Audio & video editing — podcast/video editing
Exhibit tracking slip
Verdict
Cutting a clip by deleting words from its transcript — transcribe locally, map words to timestamps, delete text to cut audio or video — is a real weekend build on open transcription tools. What doesn't come with it is Descript's actual editor: a full nonlinear timeline, AI voice cloning and overdub, and the reliability of a mature rendering pipeline.
Exhibit A — The prompt
Received on30.07.2026Build a text-based clip cutter: transcribe a dropped-in audio or video file locally with word-level timestamps (or through a hosted transcription API if a key is configured), then render the transcript as clickable words. Selecting and deleting words or whole sentences adds them to a cut list, undoable before rendering. Render by building a single ffmpeg command that trims the cut spans out of the original media, padding each cut by a small buffer so words aren't clipped mid-syllable. Save each project (transcript plus cut list) as a file next to the source media so an edit can be reopened and refined later.
Do not build a multitrack timeline, AI voice cloning or overdub, automatic filler-word detection, or any cloud features — this is a cutter, not an editor, and that's the honest, complete scope for a personal tool.Opening prefills the prompt — press enter to run it.
Exhibit B — What you lose
- B.1 a polished nonlinear timeline editor
- B.2 AI voice and overdub tools
- B.3 automatic filler-word removal
- B.4 project templates
- B.5 cloud collaboration
Prior art
Exhibit C — Why people still pay: media pipeline/collaboration
People pay to avoid building and maintaining a media editor themselves, not just for transcription — the editing and rendering pipeline is the real product.
Questions
Can I import my existing Descript projects?
No — Descript's project format is proprietary. You'd start from an exported media file and transcribe it fresh here, not carry over the editable project.
Will it work on my phone?
No — transcription needs real processing power, and text-based cutting is a desktop workflow. This tool isn't meant to run on a phone.
What does it cost to run?
Nothing if you transcribe locally — no hosting, no API key. Using a hosted transcription API instead costs a small amount per minute of audio.
What's the one thing that doesn't survive the rebuild?
The actual editor. This tool deletes words to cut a clip; it has no timeline, no multitrack audio, no AI voice tools — for anything beyond transcript-based cutting, you're back to a real video editor.
Related tools
Receipt