shorts

Orchestrate transcription, scoring, and Remotion rendering to create vertical short clips.

177|41|Updated Feb 16, 2026
One-click install
npx skills add https://github.com/AgriciDaniel/claude-shorts --skill shorts-agricidaniel
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: shorts
Source: https://github.com/AgriciDaniel/claude-shorts/tree/main
Command: npx skills add https://github.com/AgriciDaniel/claude-shorts --skill shorts-agricidaniel

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires faster-whisper, mediapipe, numpy, opencv-python, torch, jq, ffmpeg, ffprobe, and includes scripts (resource) and references (resource) components.

What problem does it solve?

Transform long-form videos into engaging short-form clips by orchestrating transcription, scoring, and rendering using Claude Code, faster-whisper, and Remotion.

Core Features & Use Cases

  • 10-step interactive pipeline that transcribes, scores, selects, snaps boundaries, and renders vertical shorts.
  • 3 caption styles (Bold, Bounce, Clean) rendered via Remotion for platform-ready exports.
  • GPU-accelerated transcription, content detection, and per-segment export to major social platforms.

Quick Start

Invoke the shorts skill in Claude Code with a video file to start the 10-step pipeline.

Frequently Asked Questions about shorts

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I convert long videos into vertical shorts for TikTok and Instagram Reels?

To convert long videos into vertical shorts, this Skill uses a 10-step interactive pipeline that transcribes, scores, and snaps boundaries before rendering 1080x1920 clips. It orchestrates faster-whisper and Remotion to export platform-ready videos.

Can I use Remotion and faster-whisper together to automate video captioning?

Yes, you can use Remotion and faster-whisper together to automate video captioning. The pipeline uses faster-whisper for GPU-accelerated transcription and Remotion to render captions in Bold, Bounce, or Clean styles for your shorts.

What's the best way to find viral clip boundaries in a long video?

Finding viral clip boundaries is handled by the pipeline's scoring and boundary snapping steps. It analyzes transcribed segments and content detection data to identify high-engagement moments before cutting the video.

Do I need FFmpeg and mediapipe installed to process and render video shorts?

Yes, you need FFmpeg and mediapipe installed along with faster-whisper, numpy, and torch. These dependencies handle media processing, content detection, and local 1080x1920 rendering required by the pipeline.

Does this pipeline support GPU acceleration for transcription and content detection?

Yes, the pipeline supports GPU acceleration for transcription and content detection. It relies on torch and faster-whisper to process long-form videos and detect segments faster during the interactive workflow.

Why use Claude Code to orchestrate a short-form video rendering pipeline?

Using Claude Code to orchestrate short-form video rendering allows interactive control over the 10-step pipeline. It coordinates transcription, scoring, and Remotion rendering to produce platform exports without manual scripting.