YouTube

AI video editing
for YouTube creators

For most creators the edit is the bottleneck, not the camera. EditAssist builds the rough cut from your script, finds the B-roll when you describe it, strips the filler, and spins up a vertical Short before you've finished your coffee. You direct in plain English, the agent does the timeline work, and your footage stays on your machine.

01

The creator workflow

Dump the day's footage into a folder. The agent ingests it, transcribes the talking-head, and indexes every frame so the whole shoot is searchable. Hand it your script and it assembles a rough cut against your A-roll, line by line and take by take, then tightens it so filler words come out, jump cuts smooth over and dead air goes.

When you need a shot, just describe it. Ask for “the clip where I unbox the thing” or “skyline at golden hour” and the agent pulls it and drops it on the timeline. When the main video is locked, ask for a vertical recut and it reframes to 9:16 with you kept in frame, ready to publish as a Short. A single session covers the main upload and the clips.

02

Prompts for YouTube creators

  • Generate a rough cut from the script I just uploaded and lay it against the talking-head footage.

  • Find all my reaction-cam clips where my jaw drops and cut them in after each reveal.

  • Pull every B-roll shot of the city skyline and arrange them by time of day.

  • Find all the moments where I say 'um' or 'uh' and remove them automatically.

  • Cut every jump cut from my last vlog and smooth the gaps with cutaways.

  • Create a vertical 9:16 recut of this horizontal sequence with the subject tracked.

  • Chapter this tutorial into sections automatically and export markers for YouTube.

  • Cut a 30-second teaser from the full video and optimise it for YouTube Shorts.

Hundreds more in the prompt library.

03

What powers a creator's edit

Semantic B-roll search

Every frame of your footage is indexed. Describe the shot, say 'the city skyline at dusk' or 'me holding up the product', and the agent finds it instantly, even across a full archive.

Rough cut from your script

Upload the script or VO and the agent assembles an editing-ready timeline against your talking-head footage, matching each line to the take that delivers it.

Auto-reframe to vertical

Turn the horizontal master into a 9:16 cut with the subject kept in frame. That gives you a Short to publish alongside the main video, with no manual keyframing.

Tighten the talk

Remove filler words and dead air, cut the jump cuts, and detect off-topic tangents. The agent does the listening so the upload feels sharp.

04

Fast turnaround, local, and free to start

The edit runs on your own GPU. Transcription, frame indexing, search and reframing all happen locally, so there's no upload wait and your unreleased footage never leaves your machine. EditAssist is free to download and the local models are free to run, and new accounts get £15 of credit with no card. If you lean on cloud models, pay as you go, or pick £20/mo Starter or £100/mo Pro for included credit. The agent operates DaVinci Resolve, Premiere Pro, After Effects and Final Cut directly, whatever you already cut in.

05

Getting started

  1. 01Download EditAssist for macOS or Windows and sign up. You get £15 credit and no card is needed.
  2. 02Point it at your shoot folder, and it ingests, transcribes and indexes every frame.
  3. 03Upload your script and ask for a rough cut, then refine and pull B-roll in plain English.
  4. 04Lock the main video, ask for a 9:16 Short, and deliver both in one go.

More workflows: podcast editing, social clips, documentary editing. Or see everything EditAssist connects to and browse the prompt library.