Azurely 🎙️

Transform your meeting recordings into actionable summaries — powered by Azure AI

---

📌 Overview

Azurely is an AI-powered meeting summarizer that takes audio recordings and returns structured summaries, key points, and action items — all in seconds.

Upload an audio file → get a full transcription + executive summary + action items. Built with Azure Speech Service for transcription and Azure OpenAI for intelligent analysis.

---

🚀 Features

• 🎧 Multi-format audio support — MP3, WAV, M4A, OGG, MP4
• 📝 Full transcription — powered by Azure Speech SDK with continuous recognition
• 🤖 AI analysis — executive summary, key points, and action items via GPT-4o-mini
• ⚡ Chunk processing — handles long recordings by splitting into 2-minute segments
• 🌍 Multilingual — supports es-ES, es-MX, en-US, en-GB, pt-BR
• 🔒 Auto cleanup — temporary files deleted after every request

---

🏗️ Architecture

Audio Upload (MP3/WAV/M4A)
        ↓
  Format Validation
        ↓
  ffmpeg Conversion
  (→ WAV 16kHz mono)
        ↓
  Audio Chunking
  (2-min segments)
        ↓
  Azure Speech SDK
  (Transcription)
        ↓
  Azure OpenAI
  (GPT-4o-mini Analysis)
        ↓
  JSON Response
  (Summary + Action Items)

---

📁 Project Structure

azurely/
├── backend/
│   ├── app/
│   │   ├── main.py                # FastAPI app, CORS, lifespan
│   │   ├── core/
│   │   │   ├── config.py          # pydantic-settings
│   │   │   └── exceptions.py      # custom HTTP exceptions
│   │   ├── models/
│   │   │   ├── meeting.py         # MeetingSummary, ActionItem
│   │   │   └── requests.py        # AnalyzeRequest
│   │   ├── services/
│   │   │   ├── speech.py          # Azure Speech STT
│   │   │   └── openai_service.py  # Azure OpenAI analysis
│   │   ├── utils/
│   │   │   ├── audio.py           # ffmpeg conversion + chunking
│   │   │   └── cleanup.py         # temp file cleanup
│   │   └── api/
│   │       ├── dependencies.py    # file validation
│   │       └── routes/
│   │           ├── health.py      # GET /health
│   │           └── analyze.py     # POST /api/analyze
│   ├── .env.example
│   ├── requirements.txt
│   └── requirements-dev.txt
└── frontend/
    ├── index.html
    └── assets/
        ├── css/styles.css
        └── js/
            ├── api.js
            ├── ui.js
            └── main.js

---

⚙️ Setup

Prerequisites

• Python 3.13+
• ffmpeg installed and in PATH
• Azure account with:
  • Azure Speech Service resource
  • Azure OpenAI resource with a deployed model

Installation

# Clone the repo
git clone https://github.com/AndresRJ18/Azurely.git
cd Azurely/backend

# Create virtual environment
python -m venv venv
source venv/Scripts/activate  # Windows
# source venv/bin/activate    # Linux/Mac

# Install dependencies
pip install -r requirements-dev.txt

# Configure environment
cp .env.example .env
# Edit .env with your Azure credentials

Environment Variables

# .env
AZURE_SPEECH_KEY=your_speech_key
AZURE_SPEECH_REGION=eastus
AZURE_OPENAI_ENDPOINT=https://your-resource.openai.azure.com/openai/v1
AZURE_OPENAI_KEY=your_openai_key
AZURE_OPENAI_DEPLOYMENT=gpt-4o-mini
MAX_AUDIO_SIZE_MB=25
TEMP_DIR=/tmp/azurely

Run

python -m uvicorn app.main:app --reload --port 8000

API available at http://localhost:8000 Swagger UI at http://localhost:8000/docs

---

📡 API Reference

GET /health

Verifica conectividad con Azure Speech y Azure OpenAI.

{
  "status": "ok",
  "service": "Azurely API",
  "azure_speech": "connected",
  "azure_openai": "connected"
}

POST /api/analyze

Analiza un archivo de audio y retorna el resumen.

Request — multipart/form-data

Field	Type	Required	Default
file	audio file	✅	—
language	string (BCP-47)	❌	es-ES

Response

{
  "summary": "Resumen ejecutivo de la reunión...",
  "key_points": [
    "Punto clave 1",
    "Punto clave 2"
  ],
  "action_items": [
    {
      "task": "Enviar el plan de medios",
      "assignee": "Carlos",
      "deadline": "viernes"
    }
  ],
  "transcription": "Transcripción completa del audio...",
  "language_detected": "es-ES",
  "duration_estimate": "5 minutos"
}

Supported languages

Code	Language
es-ES	Español (España)
es-MX	Español (México)
en-US	English (US)
en-GB	English (UK)
pt-BR	Português (Brasil)

---

🛠️ Tech Stack

Layer	Technology
Backend framework	FastAPI
Speech-to-Text	Azure Speech SDK
AI Analysis	Azure OpenAI (GPT-4o-mini)
Audio processing	ffmpeg + ffmpeg-python
Config management	pydantic-settings
Async file I/O	aiofiles

---

👥 Team

Built at a hackathon by a team of 4 Informatics Engineering students.

Role	Scope
Backend	FastAPI, Azure services, audio pipeline
Frontend	UI, file upload, results rendering
DevOps	Docker, Azure Container Apps deployment

---

📄 License

MIT License — see LICENSE for details.

---

_{Built with ❤️ using Azure AI Services}