ReelMind V3

ReelMind is a video processing automation tool that extracts clips from YouTube videos and formats them for short-form video platforms like TikTok, YouTube Shorts, and Instagram Reels. It utilizes AI for transcript analysis and local FFmpeg processes for video rendering and captioning.

---

Features

• Video Analysis: Uses Gemini via OpenRouter to analyze YouTube .json3 subtitle transcripts. The AI identifies 15-60 second segments based on a predefined engagement scoring prompt.
• Automated Captions: Extracts audio and processes it using the OpenAI Whisper API to generate word-level timestamps. These timestamps are converted into .ass subtitle files and burned into the video.
• Local Rendering Pipeline: Utilizes local FFmpeg to process videos without relying on cloud rendering services. Supports cropping to a 9:16 aspect ratio and provides a web interface for manual adjustment of start and end times.
• Background Audio Mixing: Supports mixing an optional background audio track (public/bgm.mp3) underneath the primary video audio.
• Split-Screen Compositing: Includes an option to overlay the primary video on top of secondary background footage (e.g., gameplay videos) to create a vertically stacked layout.
• Publishing Integrations:
  • YouTube Shorts: Integrates with the official googleapis OAuth for direct uploading.
  • TikTok: Includes a headless Microsoft Playwright script (scripts/tiktok-bot.js) that uses session cookies to automate the upload process.

Getting Started

Prerequisites

Ensure you have the following installed on your local machine before running ReelMind:

• Node.js (v18 or higher)
• FFmpeg (Must be added to your system PATH)
• yt-dlp (Must be added to your system PATH)
• Playwright Browsers (Run npx playwright install after npm install)

Installation

1. Clone the repository

   git clone https://github.com/obzennn/ReelMind.git
   cd ReelMind

2. Install dependencies

   npm install
   npx playwright install chromium

3. Configure the Environment Create a .env.local file and insert your required API keys:

   # Required: For Video Transcript Analysis (OpenRouter)
   OPENROUTER_API_KEY="sk-or-v1-YOUR-KEY-HERE"
   
   # Required: For Whisper API Audio Transcription
   OPENAI_API_KEY="sk-proj-YOUR-KEY-HERE"
   
   # Optional: For YouTube Shorts Auto-Publish
   YOUTUBE_CLIENT_ID="your-client-id"
   YOUTUBE_CLIENT_SECRET="your-client-secret"
   YOUTUBE_REDIRECT_URI="http://localhost:3000"
   YOUTUBE_REFRESH_TOKEN="your-refresh-token"

4. TikTok Automation Setup If you plan to use the TikTok publisher script, you need to generate a session cookie:

   • Edit scripts/tiktok-bot.js and temporarily set headless: false.
   • Run the script manually: node scripts/tiktok-bot.js ./test.mp4 "Title" "Desc" "#shorts"
   • Scan the TikTok login QR code in the browser window.
   • Close the browser. The session cookie is saved in .tiktok_session. You can restore headless: true.

5. Run the development server

   npm run dev

   Open http://localhost:3000 with your browser to access the interface.

Architecture

• Frontend: Next.js 14 (App Router) with Tailwind CSS.
• Backend API: Next.js Route Handlers.
• Video Engine: Uses yt-dlp-exec, execa, and FFmpeg (-filter_complex for stacking and amix for audio routing).
• Automation: Playwright for browser automation and Google APIs for OAuth integrations.

---

Disclaimer: ReelMind is an independent open-source project. Users are responsible for complying with the Terms of Service of YouTube, TikTok, and other platforms, as well as respecting copyright laws when downloading and processing videos.