explaingit

cgallic/video-review-os

Analysis updated 2026-05-18

2PythonAudience · generalComplexity · 4/5Setup · hard

TLDR

A local Python pipeline that takes raw video files and automatically transcribes, clips, repairs, assembles, renders, and captions them, producing a review dashboard before anything is published.

Mindmap

mindmap
  root((Video Review OS))
    What it does
      Clip candidate selection
      Quality gate and repair
      Assembly and rendering
      Caption generation
    Outputs
      MP4 draft renders
      SRT and VTT captions
      SVG thumbnail drafts
      Review dashboard
    Tech stack
      Python
      ffmpeg
      Whisper transcription
      SVG and JSON
    Use cases
      Creator clip workflow
      Local video review
      Pre-publish approval
Click or tap to explore — scroll the page freely

Code map

Detail Auto

An interactive map of this repo's files and how they connect — its source is parsed live in your browser. Click Visualize to build it.

filefunction / class

What do people build with it?

USE CASE 1

Automatically find and clip the best moments from raw live performance or interview footage.

USE CASE 2

Build a local review pipeline where every clip draft is inspected and approved before posting.

USE CASE 3

Generate transcript-based captions and thumbnail drafts for short-form video clips without using a hosted service.

What is it built with?

PythonffmpegWhisperSVGJSON

How does it compare?

cgallic/video-review-os0-bingwu-0/live-interpreter0xkaz/llm-governance-dashboard
Stars222
LanguagePythonPythonPython
Setup difficultyhardmoderatehard
Complexity4/52/54/5
Audiencegeneralgeneralops devops

Figures from each repo's GitHub metadata at analysis time.

How do you get it running?

Difficulty · hard Time to first run · 1h+

Requires Python 3.11+, ffmpeg, and optionally a local Whisper installation, no hosted service or social accounts are needed.

License not stated in the README.

In plain English

Video Review OS is a local video clipping pipeline for creators and small teams who want full control over every step before anything gets published. You drop raw video files into a folder and the tool handles the rest: it transcribes the footage, finds candidate clip ranges, scores each clip's quality from an audience perspective, proposes ways to combine and repair clips into finished assemblies, renders MP4 drafts, writes caption files, pulls frame stills, generates thumbnail SVGs, and builds a static review dashboard. Nothing publishes automatically. Every output waits for a human to approve it first. The quality gate step is more than a keep-or-reject filter. When a clip has problems (weak ending, awkward pause, filler opening), the tool emits concrete repair instructions describing exactly what to trim or drop, then applies those repairs automatically when assembling the final edit. The assembly layer can combine multiple clip ranges into a single video, reordering them and inserting title or bridge cards as needed, based on a storyboard proposal that can come from a simple deterministic algorithm or from your own language model endpoint. All outputs are plain files: JSON sidecars, SRT and VTT caption files, JPEG frames, SVG drafts, rendered MP4 files, and static HTML. There is no hidden database. Source videos are never deleted. The tool runs entirely on your own machine using ffmpeg for video processing and Whisper (or faster-whisper) for transcription. The tool compares itself to services like Opus Clip and Descript, noting that those are better for polished hosted workflows with social publishing. Video Review OS is for teams that want to inspect every artifact before anything leaves the machine and are comfortable running a Python pipeline locally. Setup requires Python 3.11 or higher, ffmpeg, and optionally a local Whisper installation. The README is very detailed, covering the full pipeline, all output files, and configuration options. The license is not mentioned.

Copy-paste prompts

Prompt 1
I have 10 raw MP4 recordings from a conference. Walk me through setting up Video Review OS to find the best clip candidates from each video and generate a review dashboard.
Prompt 2
How does Video Review OS decide which clips to keep versus reject? What does the quality gate check for?
Prompt 3
I want Video Review OS to use a local language model instead of the deterministic storyboard fallback. How do I configure my own LLM endpoint?
Prompt 4
After Video Review OS runs, show me how to read the assemblies.json file to understand how the tool wants to combine my clips.

Frequently asked questions

What is video-review-os?

A local Python pipeline that takes raw video files and automatically transcribes, clips, repairs, assembles, renders, and captions them, producing a review dashboard before anything is published.

What language is video-review-os written in?

Mainly Python. The stack also includes Python, ffmpeg, Whisper.

What license does video-review-os use?

License not stated in the README.

How hard is video-review-os to set up?

Setup difficulty is rated hard, with roughly 1h+ to a first successful run.

Who is video-review-os for?

Mainly general.

Open on GitHub → Explain another repo

This repo across BitVibe Labs

Scan in gitsafehub Deploy in gitdeployhub cgallic on gitmyhub

Verify against the repo before relying on details.