Personal project · local GPU

Clips in.
A cut out.

MovieMakerAI is an agentic editor: it picks music, reads beats, plans shots, and renders a music-synced movie. Models run on your GPU via vLLM — nothing is uploaded to a cloud editor.

Install & run Open the repo

Pipeline

01

Prepare

Clips become one timeline. A local model recommends a mood and a CC0 track — or you drop in your own music.

02

Plan

Screenwriter → editor → reviewer agents pick timestamps against the beat. You can watch the log as they work.

03

Cut

ffmpeg assembles the movie. Web UI at localhost:7860, or the CLI pipeline if you already live in a terminal.

Install

This one needs a GPU

NVIDIA GPU with about 40GB+ VRAM (tested on H100 80GB), Linux, Python 3.12, ffmpeg. Full detail is in the README — short path:

git clone https://github.com/GumyGam/moviemakerai.git
cd moviemakerai
bash install.sh
source venv/bin/activate
python app.py

Open http://localhost:7860. CLI: bash scripts/run_pipeline.sh

Amir Gam · built on CutClaw · github.com/GumyGam/moviemakerai