Automatic Meeting Assistant is a web-based application that lets you upload an MP3 meeting recording and receive a structured, AI-driven summary. The system uses a React frontend, a Go backend with the Gin web framework, and the OpenAI Whisper & GPT APIs for transcription and summarization.
- Upload your MP3 meeting recordings for instant AI-powered summaries
- Step-by-step upload, transcribe, and summarize progress indicators
- Robust validation: only valid
.mp3files <10MB allowed - Secure handling: temporary storage only, deleted after processing
- Download meeting summaries as
.txtor.mdformats - Responsive, modern UI
- Docker and Docker Compose
- An OpenAI API Key (for Whisper and GPT APIs)
- This repository cloned to your machine
git clone https://github.com/your-username/meeting-assistant.git
cd meeting-assistantCreate a file named .env in your project root (same directory as docker-compose.yml).
It should contain your OpenAI API key:
OPENAI_API_KEY=your-openai-api-key-hereBuild the frontend + backend images:
docker builddocker-compose upThis will:
- Start the backend at http://localhost:8080
- Start the frontend at http://localhost:3000 (via Nginx proxy)
- Visit http://localhost:3000 in your browser
- Upload an .mp3 file under 10MB
- Watch the real-time progress bar as your meeting is uploaded, transcribed, and summarized
- View and download your summary as .txt or .md
- Only accepts and processes valid .mp3 files under 10MB
- All files are sanitized on the backend; invalid/malicious files are rejected
- Uploaded audio is deleted after processing
- No persistent data storage or authentication in this demo (add as needed for production)
- Frontend: React, HTML/CSS, Nginx (for static serving and proxy)
- Backend: Go, Gin
- AI APIs: OpenAI Whisper (speech-to-text), GPT (summarization)
- Deployment: Docker, Docker Compose