Transcribai - AI-Powered Social Media Insights 🚀

Transcribai is a powerful SaaS tool that transforms lengthy videos from YouTube, Instagram, LinkedIn, and TikTok into concise, actionable insights using AI.

✨ Features

• Multi-Platform Support: Fetch transcripts from YouTube, Instagram, LinkedIn, and TikTok.
• AI-Powered Insights: Uses Groq (Llama 3 70B) to extract key takeaways and summaries.
• Premium UI: Stunning dark-mode interface built with Next.js, Framer Motion, and Tailwind CSS.
• One-Click Downloads: Export insights as professionally styled PDFs or shareable Images.
• Dynamic Adaptation: The interface adapts its colors and icons based on the selected platform.

🛠️ Tech Stack

Frontend

• Next.js 15 (App Router)
• TypeScript
• Tailwind CSS v4 + Framer Motion
• Lucide React (Icons)

Backend

• FastAPI (Python)
• LangChain + Groq (AI Logic)
• YouTube Transcript API
• yt-dlp fallback (caption extraction fallback for cloud-hosted environments)
• xhtml2pdf & PyMuPDF (PDF/Image Generation)

🚀 Getting Started

Prerequisites

• Node.js & npm
• Python 3.9+
• Groq API Key

Installation

1. Clone the repository

   git clone https://github.com/USER/Transcribai.git
   cd Transcribai

2. Backend Setup

   cd backend
   pip install -r requirements.txt
   uvicorn main:app --reload
   # Backend runs on http://localhost:8000

3. Frontend Setup

   cd frontend
   npm install
   npm run dev
   # Frontend runs on http://localhost:3000

4. Environment Variables

   • Create a .env in backend/ and add:

     GROQ_API_KEY=your_groq_api_key_here
       # Recommended for production on Render (single proxy URL)
       TRANSCRIPT_PROXY_URL=http://username:password@proxy-host:port
       # Or provide separate values
       # TRANSCRIPT_PROXY_HTTP_URL=http://username:password@proxy-host:port
       # TRANSCRIPT_PROXY_HTTPS_URL=http://username:password@proxy-host:port
       # Optional: absolute path to cookies file for yt-dlp fallback (if needed)
       YTDLP_COOKIES_FILE=/path/to/cookies.txt

☁️ Render / Cloud Note

Some cloud IP ranges are blocked by YouTube for transcript scraping. The backend now tries:

1. youtube_transcript_api (primary)
2. yt-dlp captions fallback (automatic/manual subtitle tracks)

If both fail due to network/IP blocking, use a proxy-enabled setup in production.

🚀 Render Deployment (Verified Settings)

Configure two Render web services (backend + frontend) with these values:

1) Backend service (Python)

• Root Directory: backend
• Build Command: pip install -r requirements.txt
• Start Command: uvicorn main:app --host 0.0.0.0 --port $PORT
• Required Environment Variables:
  • GROQ_API_KEY=your_groq_api_key
• Recommended Environment Variables (to bypass cloud IP blocks):
  • TRANSCRIPT_PROXY_URL=http://username:password@proxy-host:port
  • or both TRANSCRIPT_PROXY_HTTP_URL=... and TRANSCRIPT_PROXY_HTTPS_URL=...
  • or (recommended to avoid encoding/auth issues):
    • TRANSCRIPT_PROXY_HOST=pr.oxylabs.io
    • TRANSCRIPT_PROXY_PORT=7777
    • TRANSCRIPT_PROXY_USERNAME=<proxy_username>
    • TRANSCRIPT_PROXY_PASSWORD=<proxy_password>
• Optional Environment Variables:
  • YTDLP_COOKIES_FILE=/opt/render/project/src/backend/cookies.txt

2) Frontend service (Next.js)

• Root Directory: frontend
• Build Command: npm ci && npm run build
• Start Command: npm run start -- -p $PORT
• Required Environment Variables:
  • NEXT_PUBLIC_API_BASE_URL=https://<your-backend-service>.onrender.com/api

Why this matters

• Frontend now reads API URL from NEXT_PUBLIC_API_BASE_URL instead of hardcoded localhost.
• Backend uses GROQ_API_KEY from environment.
• Uvicorn binds correctly to Render's dynamic port via $PORT.

📝 API Endpoints

• POST /api/transcript - Fetch transcript from URL
• POST /api/generate - Generate insights from transcript
• POST /api/download - Download insight as PDF/Image

🤝 Contributing

Contributions are welcome! Please fork the repo and submit a PR.

📄 License

MIT License © 2026 Transcribai