jarredr092699/compilify

★ 0PythonAudience · vibe coderComplexity · 2/5ActiveLicenseSetup · easy

Mindmap

mindmap
  root((compilify))
    Inputs
      Before clips
      After clips
      Optional music
    Outputs
      Vertical compilation
      Title cards
      YouTube cut
    Use Cases
      Make TikTok reels
      Build progress videos
      Auto-trim sharp frames
    Tech Stack
      Python
      ffmpeg
      Claude Code
      numpy

mindmap root((compilify)) Inputs Before clips After clips Optional music Outputs Vertical compilation Title cards YouTube cut Use Cases Make TikTok reels Build progress videos Auto-trim sharp frames Tech Stack Python ffmpeg Claude Code numpy

Things people build with this

USE CASE 1

Auto-build a before and after compilation for TikTok or Reels from two folders of raw clips

USE CASE 2

Generate a 16 by 9 YouTube cut of a project progress reel with custom title text

USE CASE 3

Sample and score clip frames with Laplacian variance to pick the sharpest trim point

USE CASE 4

Mix a background music track with a fade out into an ffmpeg-assembled compilation

Tech stack

PythonffmpegClaude Codenumpy

Getting it running

Difficulty · easy Time to first run · 30min

Needs ffmpeg, Python 3.10+, numpy, and Claude Code; copy SKILL.md plus two scripts into the skills directory.

MIT license, so you can use, modify, and redistribute it freely as long as you keep the copyright notice.

In plain English

Compilify is a small add on for Claude Code, Anthropic's command line coding assistant, that turns raw video clips into a polished before and after compilation. You drop your footage into two folders, type a slash command, and the assistant assembles a short video ready to post on TikTok, Instagram Reels, or YouTube. The workflow is built around two folders called input/before and input/after. The skill samples three frames from every clip, scores them for sharpness using a standard image processing measure called Laplacian variance, and also scores brightness. The best window in each clip becomes the trim point. Claude then proposes a running order, usually a BEFORE title card, the before clips, an AFTER title card, and the after clips, and you can confirm or rearrange before it renders. The heavy lifting is done by ffmpeg, the standard open source video tool. One Python script handles the frame scoring and another runs ffmpeg to make title cards, normalize all clips to the same resolution and frame rate, concatenate them, and optionally mix in a background music track with a fade out. The default output is vertical 9 to 16 at around 20 to 30 seconds, but you can ask for a 16 to 9 YouTube version, custom title text such as Day 1 and Day 30, a music file, or a target duration. To install, you copy SKILL.md and two Python scripts into your Claude Code skills directory and set up a workspace folder. Requirements are Claude Code, ffmpeg, Python 3.10 or newer, and numpy. The license is MIT.

Copy-paste prompts

Prompt 1

Install compilify into my Claude Code skills folder and verify the slash command works on a test workspace

Prompt 2

Run compilify on input/before and input/after to produce a 25 second vertical reel with Day 1 and Day 30 title cards

Prompt 3

Switch the output to 16 by 9 YouTube format and add my song.mp3 with a 2 second fade out

Prompt 4

Change the frame-scoring weights in the Python script so brightness counts more than sharpness for trim selection

Prompt 5

Add a third folder called input/middle and update the running order to interleave before, middle, after

Open on GitHub → Explain another repo

Generated 2026-05-22 · Model: sonnet-4-6 · Verify against the repo before relying on details.