Film it. Your agent cuts it.

My last 50-minute recording: a draft in 2 hours, a double-checked cut in 3.5. My part was three reviews.

Editing a talking-head video is not hard. It is long.

Fifty minutes of footage. The same sentence said four times. Pauses, stutters, a glance at your notes every minute. You know exactly what the cut should be. It still takes a day to get there.

The AI tools that promise to fix this cut silences by the transcript. Transcripts lie about pauses. So you get clipped words, jumpy joins, and an edit you redo by hand.

What this is

Edit Studio is the pipeline I edit my own long-form videos with, packaged for your AI agent. You do not need an editing app: it renders the finished video for you. If you have DaVinci Resolve Studio, it uses it, and every edit lands as a real timeline you can open, change and undo. If you work in Premiere Pro, you get the cut as a timeline you import.

A tutorial shows you one graphic. This is the whole edit.

Free tutorials on making motion graphics with Claude are everywhere. They show you one graphic. Then you are on your own: find the workflow, wire the tools together, and fix it yourself the day Claude, Resolve or Premiere change.

Edit Studio is the workflow, already wired and written down, from raw footage to the finished video. Step-by-step guides take you through the install, DaVinci Resolve and Premiere Pro. And it does not go stale.

Lifetime updates

Buy once and every new version is yours, free. When I find a better way to edit my own videos, it goes into the kit: new templates and graphic styles, new steps, fixes for new versions of Resolve, Premiere and your agent. You get an email, and the new zip is waiting on your Start here page. You never go back to searching for tutorials.

Next on the list: graphic styles beyond the paper look, and a hero-object template for your first seconds.

How it works

1. Sync and transcribe. Your separate mic is locked to every camera clip, clock drift included. A word-level transcript keeps every retake.

2. Pick the takes. One agent per section keeps the last complete take of every line and removes the mistakes inside it. Anything it is unsure about becomes a marker for your ear.

3. You approve the draft. A draft and a review sheet of every cut. Nothing is committed without your yes.

4. Tighten. Dead air is judged by sound, not by the transcript. Every join follows one rule: cut after your voice decays, start on the next word's first strong frame. Audio cross fades, clean picture cuts.

5. It checks itself. A lost-words audit finds the quiet words a pause trimmer eats, often numbers and "not". Then fresh agents re-listen to every join.

6. You approve the cut. Fix a few joins by hand and the system measures your edits and applies the same rule to the rest.

7. Motion graphics (optional). Templates rendered as transparent overlays over your real cut, with their own sound effects. Every stretch where you read your notes gets a full-screen graphic. You vote on each one on a review page.

Graphics from three of my own episodes, re-rendered in English with the same kit: a few from its 12 templates, most written by my agent from its blank scaffold.

Built for

One person talking to a camera. YouTube essays, tutorials, explainers, courses. A Mac with Apple silicon or a Windows 10/11 PC, and an AI agent like Claude Code, Codex or Cursor. DaVinci Resolve Studio or Premiere Pro if you have one; no editing app needed.

Not for

Multi-camera podcasts and two-person interviews. Anyone who wants a one-click button: this is a pipeline with three decisions that stay yours.

One outsourced edit of a 15-minute video costs more than this. You keep it, and every update to it, for every video after.