Repo Explainer Video Generator

Generate stunning 30-second explainer videos from any git repository using Gemini VEO 3.1

Overview

This skill analyzes a git repository (local or remote) and automatically creates a professional 30-second explainer video featuring:

• Clean, modern graphics with dynamic infographics
• AI-generated voiceover narration
• Lo-fi/trip hop background music
• Smooth scene transitions
• 1080p HD output

How It Works

1. Repository Analysis: Uses parallel AI agents to comprehensively analyze README, documentation, and code structure
2. Script Generation: Creates a 5-scene narrative script (6 seconds per scene)
3. Visual Generation: Generates 5 video clips in parallel using Gemini VEO 3.1 Fast with scene extension
4. Voiceover: Creates natural AI narration using ElevenLabs
5. Composition: Combines video, voiceover, and music using ffmpeg

Usage

/repo-explainer <path-or-url>

Examples

# Local repository
/repo-explainer /Users/you/projects/my-app

# Remote GitHub repository
/repo-explainer https://github.com/username/repo

# With custom output path
/repo-explainer https://github.com/vercel/next.js --output ./nextjs-explainer.mp4

Requirements

Environment Variables

Required API keys (set in your .env or environment):

# Required
GEMINI_API_KEY=your_gemini_api_key_here
ELEVENLABS_API_KEY=your_elevenlabs_api_key_here

# Optional
OUTPUT_DIR=./videos  # Default: current directory
VIDEO_QUALITY=1080p  # Options: 720p, 1080p (default)
VEO_MODEL=fast       # Options: fast (default), standard

System Dependencies

• Python 3.9+
• ffmpeg (for video composition)

Python Dependencies

See requirements.txt:

• google-genai
• elevenlabs
• gitpython
• moviepy
• asyncio

Output

Creates an MP4 file with:

• Resolution: 1080p (default) or 720p
• Duration: ~30 seconds
• Audio: Voiceover + background music
• Format: H.264 video, AAC audio

Scene Structure

The generated video follows this 5-scene structure (6 seconds each):

1. Hook (0-6s): Eye-catching intro with the tool's name and tagline
2. Problem (6-12s): What problem does this tool solve?
3. Solution (12-18s): How does it work? Architecture overview
4. Key Features (18-24s): Highlight 2-3 standout features
5. Get Started (24-30s): Quick start guide and call-to-action

Architecture

• Ultra-parallel processing: Repository analysis, video generation, and voiceover creation all run in parallel
• Scene extension: VEO clips seamlessly connect using the final frame of the previous clip
• Cost-optimized: Uses VEO 3.1 Fast ($0.15/sec = ~$4.50 per video) for fast, affordable generation
• Quality-first: ElevenLabs voiceover for natural, engaging narration

Prompting Strategy

The skill uses Gemini 2.5 Flash to enrich prompts with:

• Cinematic camera movements
• Specific visual styles (modern tech + infographic aesthetics)
• Detailed scene descriptions
• Synchronized audio cues

Cost Estimate

Per 30-second video:

• VEO 3.1 Fast: 30 seconds @ $0.15/sec = $4.50
• Gemini 2.5 Flash analysis: ~$0.05
• ElevenLabs voiceover: ~$0.01 (with free tier)
• Total: ~$4.50-5.00 per video

Notes

• First run may take 3-5 minutes as clips are generated
• Supports both public and private repositories (if you have access)
• Automatically clones remote repos to temp directory
• Generates intermediate files in .repo-explainer-cache/ (can be deleted after)

Credits

Built with:

• Gemini VEO 3.1 - Video generation
• ElevenLabs - AI voiceover
• Claude Code - Skill framework