Caption: Audio/Video Transcription Platform

Overview

Caption is a full-stack application for transcribing audio and video files with word-level timestamps and confidence scores.
It consists of:

• Model Server: FastAPI backend (with ML models, e.g., Whisper) running in Docker.
• Client: React frontend for uploading files, live recording, and managing transcriptions.

---

Features

• Upload audio/video files for transcription
• Live audio recording and transcription
• Word-level timestamps and confidence scores
• User authentication (Supabase)
• Stripe integration for credits/payments
• Background processing (optionally with Celery + Redis)
• Download SRT and JSON results
• Responsive UI

---

Project Structure

Caption/
│
├── model-service/         # FastAPI backend (Dockerized)
│   ├── api/               # API routers (audio, video, etc.)
│   ├── handlers/          # Processing logic (e.g., VideoProcessor)
│   ├── main.py            # FastAPI app entrypoint
│   ├── requirements.txt   # Python dependencies
│   ├── Dockerfile         # Docker build file
│   └── ...                # Other backend files
│
├── client/                # React frontend
│   ├── src/
│   │   ├── components/    # React components (upload, audiototext, record, etc.)
│   │   └── ...            # Other frontend code
│   └── ...                # React config, public, etc.
│
├── docker-compose.yml     # (If using Celery/Redis)
└── README.md

---

Getting Started

Prerequisites

• Docker
• Node.js (for client)
• Supabase project (for auth)
• (Optional) Redis and Celery for background tasks

---

1. Backend: Model Server

Build and Run with Docker

cd model-service
docker build -t caption-backend .
docker run -d -p 8000:8000 --env-file .env --name caption-backend caption-backend

• The API will be available at http://localhost:8000/docs

Environment Variables

Create a .env file in model-service/ with:

SUPABASE_URL=your_supabase_url
SUPABASE_KEY=your_supabase_key
STRIPE_SECRET_KEY=your_stripe_key
# (Add any other needed variables)

API Endpoints

• POST /analyze/audio/transcribe — Transcribe audio file
• POST /analyze/video/transcribe — Transcribe video file
• GET /analyze/audio/task-status/{task_id} — Check audio task status (if using Celery)
• GET /analyze/video/task-status/{task_id} — Check video task status (if using Celery)

---

2. Frontend: React Client

Setup

cd client
npm install

Environment Variables

Create a .env file in client/ with:

VITE_SUPABASE_URL=your_supabase_url
VITE_SUPABASE_ANON_KEY=your_supabase_anon_key
VITE_EMAILJS_SERVICE_ID=your_emailjs_service_id
VITE_EMAILJS_TEMPLATE_ID=your_emailjs_template_id
VITE_EMAILJS_PUBLIC_KEY=your_emailjs_public_key

Run the Client

npm run dev

• The app will be available at http://localhost:5173 (or as shown in your terminal).

---

Optional: Background Processing with Celery + Redis

If you want to enable background task processing:

1. Add celery and redis to requirements.txt
2. Use the provided docker-compose.yml to start backend, celery worker, and redis:

docker-compose up --build

---

Usage

• Go to the client app, sign in, and upload or record audio/video.
• The backend will process the file and return results.
• Download SRT or JSON as needed.

---

Contact

• Email: hello@dobedosoft.com
• Phone: +91 9726181166
• Address: DOBODO soft, MOTA VARACHA, Surat, India
• LinkedIn
• GitHub

---

License

MIT License

---

Credits

• OpenAI Whisper
• Supabase
• Stripe
• EmailJS