categorized attachments and sender visualisations
Mindmap features
voice assisant integrated
homepage for login
Dashboard for emails with one line summary and smart search op
dashboard of downloadable attachments and related statistics
one line summary and detailed summary and entire category deletion
visualisation by category and subsciptions and easy access for high priority mails
Keywords highlighted using NER
mail sent via crunch ai voice assistant

CrunchMail

Inspiration

The inspiration for Gmail Smart Labeler came from a universal pain point we observed in professional life - email overload. As students and professionals ourselves, we noticed how much time people spend manually organizing, categorizing, and responding to emails. The statistics were eye-opening: professionals spend nearly one-third of their workweek managing emails instead of focusing on meaningful work. This realization sparked our mission to leverage AI to automate these repetitive tasks.

Due to resource constraints, we are not able to deploy this application, we needed 2 gb cloud RAM which is not available for free

What We Learned

Throughout this project, we gained valuable insights and skills:

Deep understanding of Google's OAuth2 authentication and Gmail API integration
Hands-on experience with Google's Gemini AI for natural language processing
Implementation of semantic search using FAISS and sentence transformers
Voice recognition and text-to-speech integration
Building scalable Flask applications with proper session management
The importance of rate limiting when working with AI APIs
Handling complex email MIME types and attachments

Features

Auto-categorizes emails using Gemini AI
Creates and manages intelligent label hierarchies
Detects email priority and importance automatically
Filters spam, newsletters, and promotions
Voice commands for email operations
AI-generated smart replies
Template management
Semantic search across emails
Advanced filtering options
Attachment content search
Quick preview capabilities
Email volume trends
Interaction patterns
Productivity metrics
Secure OAuth2 authentication
Gmail API integration

How We Built It

Our development process followed these key stages:

Foundation Setup
- Built a Flask web application with secure Google OAuth2 integration
- Implemented core Gmail API functionality for email access and manipulation
AI Integration
- Integrated Google's Gemini AI for intelligent email analysis
- Implemented semantic search using FAISS and sentence transformers
- Created smart labeling algorithms based on email content and patterns
Feature Enhancement
- Added voice command capabilities using speech recognition
- Implemented smart reply suggestions
- Developed attachment management system
- Created an intuitive dashboard for email analytics
User Experience
- Designed a clean, responsive web interface
- Implemented real-time label updates
- Added search functionality with semantic understanding
- Created detailed email summaries and previews

Challenges We Faced

Technical Challenges
- Managing rate limits with the Gemini API while maintaining responsiveness
- Handling various email formats and MIME types correctly
- Implementing efficient semantic search for large email volumes
- Ensuring secure OAuth2 flow and token management
AI Integration Challenges
- Training the AI to provide consistent and relevant email categorization
- Optimizing the balance between accuracy and processing speed
- Handling edge cases in email content analysis
- Managing API costs while maintaining functionality
Scalability Challenges
- Designing the system to handle large volumes of emails
- Implementing efficient caching mechanisms
- Managing session state across multiple users
- Optimizing database queries and search operations
UX Challenges
- Creating an intuitive interface for complex email operations
- Implementing real-time updates without overwhelming the server
- Ensuring smooth voice command integration
- Maintaining responsive performance with large datasets