✅ 📌 Project Name:

MinuteMate : Your AI buddy for perfect meeting minutes

---

✅ 🎯 Problem Solved

• User Google Meet ya Zoom mein live baat kar rahe hain.

• Flask web app mein login karke woh meeting browser tab ke audio ko screen share permission ke saath record karenge.

• Recording stop hone par:

  1. Audio server par save hoga.
  2. Audio → Text: Speech-to-Text API se transcript.
  3. Transcript → Gemini AI se clean summary + action items.
  4. DB mein meeting record store hoga: show, delete, download options.

---

✅ ⚙️ Tech Stack (Layer-wise)

Layer	Stack
🌐 Frontend	HTML, Tailwind CSS/Bootstrap, Vanilla JS, Jinja2 (Flask)
🎙️ Audio	JS Web API: getDisplayMedia + MediaRecorder
🐍 Backend	Python Flask 3.x, Flask-Login, Flask-SQLAlchemy
🔗 Auth	Flask-Login + hashed passwords
🗣️ STT	Google Cloud Speech-to-Text OR Whisper (local)
🧠 AI	Gemini AI (google-generativeai Python SDK)
🗄️ Database	MySQL (users, meetings, files)
🗃️ Storage	Audio files: /recordings/ folder on server
📄 Export	.txt or .pdf with pdfkit (optional)
📦 Env	.env for API keys (Gemini, Google STT)

---

✅ 📂 Folder Structure

meeting-notes-app/
│
├── app.py               # Flask entry point
├── models.py            # SQLAlchemy DB models
├── auth.py              # Login/signup logic
├── templates/           # Jinja2 templates: login, dashboard, detail
├── static/              # JS (audio capture), CSS
├── recordings/          # Uploaded meeting audio files
├── transcripts/         # Saved transcript text
├── .env                 # API keys
├── requirements.txt
└── README.md

---

✅ 📜 Database Schema — MySQL

🔑 Table: users

Column Name	Type	Details
id	INT, PK, AUTO_INC
name	VARCHAR(255)
email	VARCHAR(255), UNIQUE
password_hash	VARCHAR(255)	Hashed password
created_at	TIMESTAMP	DEFAULT CURRENT_TIMESTAMP

---

🔑 Table: meetings

Column Name	Type	Details
id	INT, PK, AUTO_INC
user_id	INT, FK → users.id
title	VARCHAR(255)	“Team Sync - 2025-06-30”
audio_file	VARCHAR(255)	Path to saved audio file
transcript	LONGTEXT	Full transcribed text
summary	TEXT	Gemini summary
action_items	JSON	Gemini action items as JSON
created_at	TIMESTAMP	DEFAULT CURRENT_TIMESTAMP

---

✅ 🗂️ Key Modules Breakdown

✅ 1️⃣ User Authentication

• auth.py

  • /signup: Name, Email, Password → hashed store.
  • /login: Email + Password check.
  • /logout: Session clear.

• Flask-Login manages session.

• Protect routes: @login_required.

---

✅ 2️⃣ Meeting Dashboard

• /dashboard:

  • Show list of old meetings: | Title | Date | View | Download | Delete |
  • “New Meeting Note” button → triggers capture.

---

✅ 3️⃣ New Meeting Flow

Step 1: Click “Start New Meeting Note”. Step 2: JS calls:

navigator.mediaDevices.getDisplayMedia({ audio: true, video: true })

User selects tab/window → mic audio shared.

Step 3: MediaRecorder records audio:

const recorder = new MediaRecorder(stream, { mimeType: 'audio/webm' });
recorder.ondataavailable = (e) => sendChunk(e.data);

Step 4: Stop:

• Stop button → all chunks merge → send to Flask /upload.

---

✅ 4️⃣ Audio Upload Endpoint

• /upload_audio: POST route

  • Receives audio blob.
  • Saves file to /recordings/meeting-xyz.webm.

---

✅ 5️⃣ Transcription Module

• Flask:

  • Pass saved audio to Google Speech-to-Text API OR
  • Local Whisper model (whisper Python lib).

• Result: Plain text transcript.

---

✅ 6️⃣ Gemini Summarization

• Gemini API call:

  You are a smart meeting assistant.
  Given these meeting notes:
  """ ... """
  Tasks:
  1. Write a short summary (<=100 words).
  2. List clear action items (Who, What, When).
  

• Save:

  • Summary → DB field.
  • Action Items → JSON column.

---

✅ 7️⃣ Meeting Viewer

• /meeting/<id>: Show page

  • Title
  • Summary
  • Action Items (formatted list)
  • Link to download .txt or .pdf.

---

✅ 8️⃣ Delete Meeting

• /delete/<id>: Deletes:

  • DB record.
  • Audio/transcript files.

---

✅ 📊 Flow Diagram

[Browser]
    ↓ Share Tab + Mic (getDisplayMedia)
    ↓ MediaRecorder → audio blob
    ↓ POST /upload_audio

[Flask Server]
    → Save file
    → Transcribe (STT)
    → Prompt Gemini
    → Store summary + actions in DB
    → Show on dashboard

---

✅ 🔐 Security Notes

• Store audio files safely, cleanup old files.
• Hash passwords.
• Use HTTPS when deploying.

---

✅ 📜 Example Requirements.txt

Flask
Flask-Login
Flask-SQLAlchemy
python-dotenv
google-generativeai
google-cloud-speech
whisper
mysqlclient
pdfkit (optional)