You filmed an event. While filming, you commented out loud on what you want done in the edit: "keep this", "that's good", "cut this", "redo". These instructions, aligned with the video's timecode, can be turned into a preliminary cut — an EDL file you import into Premiere, DaVinci or Final Cut, pre-filling the timeline with the right segments.

The rough cut is editing's most thankless job: eliminating the bad takes, identifying the segments to keep, positioning the timecodes. If you've done that work in front of a screen, you know it's three hours for a 30-minute video. With timecoded voice instructions, the AI agent does that scouting by reading the transcript.

Prerequisites

Before you start, you need:

1. A timecoded transcript of your video, in SRT or VTT format. Generate it from your recording with Whisper, CapCut, or any transcription tool. If you want to dictate your editing instructions by voice — saying "cut there" at the exact moment — a timecoded video mode is planned in KontexVoX (see below).
2. An AI agent with file-system access: Claude Code, Codex CLI, or Gemini. Not a web chatbot.
3. The transcript exported to your computer.
A rough cut is a technical file: an EDL with precise timecodes. A web chatbot doesn't generate that kind of file in your working folder. You need an agent that reads the transcript and writes the file locally.

The timecoded video mode (coming)

KontexVoX today enriches photos: you look at an image, you speak, the description is engraved into the EXIF. For video, the principle is the same but the medium changes: you watch your recording, you dictate your editing instructions as they come to you, and the app generates a timecoded file (SRT or JSON) linked to the video. This mode is in development. In the meantime, transcribe your video with any tool (Whisper, CapCut, Premiere) and run the prompt below.

What you dictated: an example

Here's what the SRT transcript of a 20-minute video contains:

00:01:12,400 → "Keep this, it's the intro." 00:03:45,200 → "Cut, I stuttered." 00:05:30,800 → "This is the highlight, keep the wide shot." 00:08:20,000 → "Failed, starting over. Cut." 00:12:15,600 → "That's good, keep all of it."

These instructions, you said while watching the video. The AI agent will parse the transcript, identify the segments to keep and to cut, and generate an EDL with precise timecodes.

The prompt

Read the timecoded transcript file (SRT or VTT format) in this folder. Analyze the voice comments to identify: segments to keep (marked by "keep", "that's good", "highlight"), segments to cut (marked by "cut", "failed", "redo"), and potentially interesting segments (no explicit marker). Generate a preliminary cut file in EDL (Edit Decision List) format with the start and end timecodes of each kept segment. Include a note for each segment with the corresponding voice comment.

What the AI produces with this prompt

An EDL file with the precise timecodes of the segments to keep. Import it into Premiere Pro, DaVinci Resolve, or Final Cut. The timeline builds itself with the right segments. The rough cut is done. You start the final edit from there instead of from scratch.

The difference from the EDL for Premiere/DaVinci prompt: this one generates a complete preliminary cut (kept/cut segments), while the other generates an EDL of point instructions (cut there, keep that). They complement each other.

Recommended tool: Claude Code

Claude Code is particularly well suited to reading a timecoded transcript and generating an EDL file or markers. It understands editing instructions in natural language and translates them into precise timecodes.

Frequently asked questions

Does the preliminary cut replace the edit?

No. It's a starting point. The AI identifies the segments to keep from your voice instructions, but the rhythm, transitions and audio mix remain your work. The advantage is that you start from a pre-filled timeline instead of positioning everything manually.

Does it work with an automatic transcript without annotations?

Yes, but it's less precise. Without your instructions ("keep", "cut"), the AI has to guess the interesting segments from the content. It works, but with more false positives. The ideal is to annotate by voice while watching the replay.

Can I target a vertical format (Shorts, Reels)?

Not with this prompt. The rough cut generates an EDL for a linear timeline. For vertical clips, use the Highlights extraction prompt instead.

Is the EDL compatible with Final Cut Pro?

CMX 3600 EDL is the universal standard, compatible with Premiere Pro, DaVinci Resolve, and Final Cut Pro. If you use Final Cut, additionally ask for an FCPXML file in the prompt.

Going further

Generate an EDL for Premiere or DaVinci Extract the highlights from a 3-hour live stream Which AI agent for your use case? Video page