For video work

Edit a video with an agent, and review every cut from your phone

The agent renders. You scrub, mark the frames that are wrong, and rewrite the narration line by line. It sends the next cut.

The words are on screen. Read them as text

How it goes

  1. Ask for the cut

    Say what you want, such as a 40-second cut from your screen recordings. The agent gets your words as a prompt.

  2. Watch it in the chat

    While it renders, the working line shows the step it's on. The cut arrives in the chat and plays there.

  3. Mark a frame

    Scrub to the frame, draw a box on what's wrong, and type or say the note.

  4. Mark a stretch

    Set an In and an Out to name a stretch: "cut this in half". A sound takes a mark on its waveform the same way.

  5. Change a narration line

    Open the narration, pick a cue, and change it to your exact words. The agent re-voices and re-renders.

  6. Get the next cut

    The agent gets every mark numbered, up to 20 in one reply, fixes them, and sends the next cut.

Run the agent in a project on your computer, outside a container. A container agent's video and audio arrive as links.

What it needs

A computer that stays onLinux or macOS, with whatever the agent renders with, such as ffmpeg, installed
An agent you already useIt does the editing and rendering. Vendetta carries the cuts and your marks.
A phoneThe iPhone app on TestFlight, or any browser

The guide: Review an agent's video from your phone

Install it and send the first clip

$ git clone https://github.com/vendetta-sh/vendetta.git
$ cd vendetta && bash install.sh
$ vendetta notify start
$ vendetta tui