Forem: Maxime Beaudoin

HomeWise AI - Intelligent Property Analysis Platform

Maxime Beaudoin — Sun, 24 Aug 2025 15:15:25 +0000

This is a submission for the AI Agents Challenge powered by n8n and Bright Data

What I Built

HomeWise AI is a modern, intelligent property analysis platform that transforms Zillow URLs into comprehensive, professional-grade property reports. Built specifically for first-time homebuyers, it leverages AI agents and real-time web data to provide detailed insights that help users make informed real estate decisions.

Key Features:

🏠 Instant Property Analysis: Paste any Zillow URL to get detailed property insights
🤖 AI-Powered Reports: Comprehensive analysis including market trends, financial projections, and neighborhood data
📱 Modern SPA Experience: Seamless single-page application with smooth transitions
📊 Smart Report Management: Automatic deduplication with 24-hour refresh cycles
🎯 First-Time Buyer Focus: Specialized insights and "Essential Tools & Preparations" sections
📄 Professional Output: Print-ready reports with comprehensive formatting
📲 Real-Time Monitoring: Telegram notifications for new report generation with property address

Problem Solved:

Property analysis typically requires hours of research across multiple sources. HomeWise AI consolidates this into a single, comprehensive report generated in under 45 seconds, providing first-time homebuyers with professional-grade insights without the complexity.

Demo

🌐 Live Application: https://homewise.maximebeaudoin.me/

Try it yourself:

Visit the live demo
Paste any Zillow property URL (e.g., https://www.zillow.com/homedetails/123-main-st/...)
Watch as the AI agent generates a comprehensive property analysis
Explore the generated report with market analysis, financial insights, and neighborhood data

n8n Workflow

📁 Complete n8n Workflow: GitHub Gist - HomeWise AI n8n Workflow

The complete n8n workflow JSON and setup instructions are available in the GitHub Gist, including:

Complete workflow JSON export ready for import
AI agent prompts and system instructions
Bright Data verified node configuration
Database integration setup
Sample data and response examples
Step-by-step implementation guide

Technical Implementation

Architecture Overview

HomeWise AI uses a modern, distributed architecture that separates concerns for optimal performance and maintainability:

Frontend: Next.js 15.4.7 with React 19 SPA

Modern single-page application with dynamic component loading
Seamless transitions without page reloads
Mobile-responsive design with Tailwind CSS 4
Smart navigation and URL management

Backend: n8n AI Agent Workflows

Complete automation logic handled by n8n workflows
AI agent for property analysis and report generation
Bright Data integration for real-time web scraping
Database operations for report storage and deduplication

Database: Supabase (PostgreSQL)

Property reports with timestamp-based versioning
Conversation tracking and session management
Optimized for fast retrieval and report reuse

AI Agent Configuration

System Instructions:
The AI agent is configured as a professional real estate analysis expert with specialized knowledge in:

Property valuation and market analysis
Financial projections and investment potential
Neighborhood analysis and risk assessment
First-time homebuyer guidance and preparation

Model Choice: GPT-5-nano for comprehensive analysis capabilities

Memory & Context:

Conversation-based memory for follow-up questions
Property data context from Bright Data extraction
Historical market data for trend analysis

Tools Used:

Bright Data verified node for Zillow data extraction
Database tools for report storage and retrieval
Timestamp logic for smart report reuse
Telegram integration for real-time monitoring and notifications
Error handling and user notification systems

Bright Data Verified Node

The Bright Data integration is central to HomeWise AI's functionality:

Data Extraction:

Primary Source: Zillow property listings with comprehensive details
Property Information: Price, specifications, history, photos, and descriptions
Market Data: Comparable sales, neighborhood statistics, and trend data
Location Intelligence: School districts, amenities, crime data, and transportation

Smart Usage Optimization:

24-Hour Report Lifecycle: Reuse existing reports to minimize API calls
Timestamp-Based Logic: Generate new reports only when data is outdated
Cost Control: Rate limiting (1 analysis per 5 minutes) and daily limits (10 per IP)
Error Handling: Graceful fallbacks for data extraction issues

Integration Benefits:

Verified Node Reliability: Leverages n8n's official Bright Data integration
Real-Time Data: Fresh property information for accurate analysis
Comprehensive Coverage: Multiple data points for thorough property evaluation
Scalable Architecture: Efficient resource usage with smart caching

Real-Time Monitoring System

Telegram Integration:
HomeWise AI includes a sophisticated monitoring system that sends real-time notifications via Telegram whenever a new property report is generated. This feature provides:

Instant Alerts: Immediate notification when property analysis is completed
Property Context: Includes property address and key details in the notification
System Monitoring: Real-time visibility into application usage and performance
Developer Insights: Track user engagement and system health
Silent Operation: Monitoring runs in the background without affecting user experience

Technical Implementation:

n8n Telegram Node: Integrated directly into the workflow after report generation
Conditional Execution: Only triggers for successful new report creation (not cached reports)
Property Data Extraction: Automatically includes address and relevant property information
Scalable Design: Can be easily extended to support multiple notification channels

Journey

Development Process

The development of HomeWise AI followed a systematic approach focusing on core functionality first, then optimization, and finally user experience polish:

Foundation & Integration: Established the core n8n workflows with AI Agent and Bright Data nodes, developed comprehensive property analysis prompts, implemented the database schema for report storage and versioning, and created the initial Next.js frontend with chat interface.

Advanced Features & Optimization: Built the smart deduplication system with 24-hour report lifecycle to optimize costs, implemented comprehensive error handling and user feedback mechanisms, developed print-optimized report layouts with professional formatting, and added rate limiting and cost control systems.

Modern SPA & Polish: Transformed the application into a modern single-page experience with dynamic component loading and smooth transitions, created a clean landing page with progressive disclosure, and added mobile-responsive design with accessibility features for an exceptional user experience.

Challenges Overcome

1. Cost Management with Bright Data

Challenge: Balancing comprehensive data extraction with API cost control
Solution: Implemented smart timestamp-based report reuse system that serves existing reports within 24 hours, reducing API calls by ~80%

2. Complex Report Generation

Challenge: Creating comprehensive, professional-grade property reports
Solution: Developed specialized AI agent prompts with structured output formatting, including market analysis, financial projections, and first-time buyer guidance

3. User Experience Optimization

Challenge: Making property analysis accessible and intuitive for first-time users
Solution: Built modern SPA with clean landing page, seamless transitions, and progressive disclosure of features

4. Database Performance

Challenge: Efficiently storing and retrieving large property reports
Solution: Implemented versioned report storage with smart indexing and optimized queries for fast retrieval

What I Learned

Technical Insights:

n8n's AI Agent node provides powerful automation capabilities for complex workflows
Bright Data's verified node integration simplifies reliable web scraping at scale
Modern React patterns (useTransition, dynamic imports) create exceptional user experiences
Smart caching strategies can dramatically reduce API costs while maintaining data freshness

Product Development:

Focused scope leads to better execution than feature-heavy approaches
First-time homebuyers need guidance, not just data - the "Essential Tools & Preparations" section became one of the most valuable features
Professional report formatting significantly increases perceived value and usability

AI Agent Design:

Specialized system instructions produce much better results than generic prompts
Structured output formatting is crucial for consistent, professional reports
Context management between conversation and report generation improves user experience

Future Enhancements

Multi-Property Comparison: Side-by-side analysis of multiple properties
Market Alerts: Automated notifications for price changes and market updates
Enhanced Visualizations: Charts and graphs for market trends and financial projections
API Access: Developer API for integration with other real estate tools

HomeWise AI demonstrates the power of combining n8n's AI Agent capabilities with Bright Data's reliable web scraping to create a focused, valuable solution for real estate analysis. The modern SPA architecture ensures an exceptional user experience while smart resource management keeps costs under control.

Screenshots & Demo Video

Application Screenshots

Demo Video

Complete User Journey Walkthrough
// Detect dark theme var iframe = document.getElementById('tweet-1959633141917684029-222'); if (document.body.className.includes('dark-theme')) { iframe.src = "https://platform.twitter.com/embed/Tweet.html?id=1959633141917684029&theme=dark" }
Full demonstration of the property analysis workflow from URL input to report generation

Live Demo: https://homewise.maximebeaudoin.me/
n8n Workflow: GitHub Gist

InboxInsights: AI-Powered Mood Analytics Through Email

Maxime Beaudoin — Mon, 26 May 2025 12:32:14 +0000

InboxInsights: AI-Powered Mood Tracking Through Email Analysis

This is a submission for the Postmark Challenge: Inbox Innovators.

InboxInsights: Transform your emotional well-being through intelligent email-based mood tracking and AI-powered insights.

💡 What I Built

InboxInsights is an AI-powered mood tracking platform that revolutionizes personal wellness monitoring by leveraging Postmark's inbound email parsing capabilities. Users can track their emotional well-being simply by sending emails describing their day, mood, and activities to a dedicated email address.

The platform uses n8n workflows integrated with Postmark's inbound webhooks to automatically process incoming emails, extract mood data using OpenAI's GPT-4, and store structured insights in a Supabase PostgreSQL database. Users can then visualize their mood trends, energy levels, and stress patterns through an interactive Next.js dashboard.

Key features include:

📧 Email-Based Data Ingestion: Send mood updates via email using Postmark's inbound parsing
🤖 AI-Powered Analysis: GPT-4 extracts mood scores, energy levels, stress indicators, and activities
📊 Advanced Analytics Dashboard: Multi-tab interface with Overview, Trends, and AI Insights
📈 Rich Visualizations: Interactive charts, mood heatmaps, distribution analysis, and trend tracking
🧠 AI-Powered Insights: Personalized recommendations based on mood patterns and correlations
📅 Mood Heatmap: GitHub-style calendar view showing mood patterns over 12 weeks
🛡️ Content Moderation: Automated content filtering with violation detection and blocking
🔐 Secure Authentication: User authentication and data privacy with Supabase Auth
📱 Responsive Design: Modern UI built with Next.js, Tailwind CSS, and shadcn/ui

🎯 Use Case

Mental health and emotional well-being tracking is often cumbersome, requiring users to remember to log into apps or fill out forms. InboxInsights solves this by meeting users where they already are - their email inbox.

Target Users:

Individuals seeking to track their emotional well-being without additional apps
Mental health professionals who want clients to maintain mood journals
Wellness enthusiasts interested in data-driven self-improvement
Busy professionals who prefer email-based workflows over dedicated apps

Problem Solved:

Friction in mood tracking: Traditional apps require remembering to log in and fill forms
Inconsistent data collection: Users often forget to track their mood regularly
Limited context: Simple mood ratings lack the rich context that natural language provides
Data silos: Mood data trapped in proprietary apps without export capabilities

🧪 Demo

🚀 Try InboxInsights Live: https://inboxinsights.me/

I strongly encourage you to test the full application! We've built extensive features and content that showcase the complete email-to-analytics workflow. The live demo includes rich mood data, advanced visualizations, and AI-powered insights.

Testing Instructions

Option 1: Use Demo Account (Recommended for Quick Testing)

Sign in using the demo credentials provided in the form
View existing mood data on the dashboard with interactive charts

Option 2: Create Your Own Account

Sign up with your own email address for a personalized experience
Start with a clean slate and build your own mood tracking history
Important: You must send mood emails from the same email address you used to sign up

For Both Options:

Send a test email to: b82ba9d30ef2dd7cf65016dfe8c69b37@inbound.postmarkapp.com
Email content example:

Subject: Daily Mood Update

Hi! Today was a pretty good day. I slept about 7 hours last night and woke up feeling refreshed.
My energy level is around 8/10 today. I went for a morning jog and had a productive work session.
Stress level is low, maybe 3/10. The weather is sunny and warm which always boosts my mood!

Activities: Morning jog, work meetings, lunch with friends

Watch real-time updates as your email gets processed and appears on the dashboard
Explore advanced features:
- Overview Tab: Enhanced mood statistics, today's mood, weekly trends, and recent entries timeline
- Trends Tab: Interactive mood charts, distribution analysis, and 12-week mood heatmap
- Insights Tab: AI-powered personalized recommendations and pattern analysis

Content Moderation Demo

Try sending inappropriate content to see the automated blocking system in action - you'll receive an immediate violation notification email.

Quick Video Demo

🧰 Code Repository

🔗 GitHub – InboxInsights

The complete source code is available on GitHub with comprehensive documentation, setup instructions, and deployment guides. The repository includes:

Full Next.js application with TypeScript
Supabase database schema and migrations
n8n workflow configuration for email processing
Docker setup for local development
Comprehensive README and documentation

⚙️ How I Built It

Architecture Overview

InboxInsights uses a modern, scalable architecture built around Postmark's inbound email processing:

📊 View Complete Workflow Diagram →

The architecture consists of four main layers:

Email Ingestion Layer: Postmark receives emails and triggers webhooks to n8n
Processing Layer: n8n workflows handle AI analysis and content moderation
Data Storage Layer: Supabase PostgreSQL stores structured mood data
Presentation Layer: Next.js dashboard provides real-time visualization

As shown in the diagram above, the system processes emails through two distinct paths:

Content Violation Path: Inappropriate content is blocked and triggers immediate email notifications
Clean Content Path: Valid mood data flows through AI analysis to the dashboard

Tech Stack

Frontend: Next.js 15, TypeScript, Tailwind CSS, shadcn/ui components
Backend: Supabase (PostgreSQL, Auth, Real-time subscriptions)
Email Processing: Postmark inbound webhooks + n8n automation
AI Integration: OpenAI GPT-4 for mood analysis and content moderation
Deployment: Vercel (frontend), Supabase Cloud (backend), self-hosted n8n

Implementation Process

1. 📬 Postmark Integration

Configured Postmark inbound server with webhook endpoint
Set up dedicated email address: b82ba9d30ef2dd7cf65016dfe8c69b37@inbound.postmarkapp.com
Implemented webhook payload processing in n8n

2. 🔄 n8n Workflow Design

Content Moderation: First node classifies incoming emails for inappropriate content
Conditional Processing: Blocks violating content and sends notification emails
AI Analysis: GPT-4 extracts structured mood data from email content
Database Integration: Stores processed data in Supabase with proper relationships

3. 🤖 AI-Powered Analysis
The AI agent extracts:

Mood Score (1-10): Overall emotional state
Energy Level (1-10): Physical and mental energy
Stress Level (1-10): Stress and anxiety indicators
Sleep Hours: Sleep duration and quality
Activities: Daily activities and events
Weather: Environmental factors affecting mood
Personalized Notes: Encouraging AI-generated feedback

4. 🗄️ Database Design

mood_entries: Core mood tracking data with email metadata
email_violations: Content moderation results and flags
Real-time subscriptions: Live dashboard updates

5. 🎨 Frontend Development

Interactive Charts: Recharts for mood trends and patterns
Real-time Updates: Supabase real-time subscriptions
Responsive Design: Mobile-first approach with Tailwind CSS

Postmark Features Utilized

1. 🔗 Inbound Email Parsing

Configured inbound server to receive emails at custom domain
Webhook integration delivers structured email data (sender, subject, body, metadata)
Automatic parsing of email headers and content

2. ⚡ Real-time Processing

Immediate webhook delivery enables real-time mood tracking
No polling required - instant processing of incoming emails
Reliable delivery with built-in retry mechanisms

3. 🛡️ Content Filtering Integration

Email metadata used for sender verification and spam protection
Integration with AI moderation for content policy enforcement
Automated response emails for policy violations

4. 📊 Analytics & Monitoring

Email processing metrics tracked in dashboard
Delivery status monitoring for outbound notifications
Error handling and logging for failed processing

🔄 Key Features & Innovations

Email-First Approach

Unlike traditional mood tracking apps, InboxInsights meets users in their existing workflow - email. This reduces friction and increases consistency in mood tracking.

AI-Powered Insights

The platform doesn't just store data - it provides intelligent analysis, extracting multiple dimensions of wellness from natural language descriptions.

Content Moderation

Built-in AI content moderation ensures the platform maintains appropriate communication standards while protecting user experience.

Advanced Analytics Dashboard

The platform features a sophisticated three-tab dashboard interface:

Overview Tab: Real-time mood statistics, today's mood tracking, weekly trends, and timeline of recent entries
Trends Tab: Interactive mood charts with time-series analysis, mood distribution pie charts, and a GitHub-style mood heatmap showing 12 weeks of data
Insights Tab: AI-powered personalized insights including mood pattern analysis, sleep correlation, stress indicators, and actionable recommendations

Real-time Dashboard

Live updates create an engaging experience where users can immediately see their mood data visualized after sending an email.

💡 Lessons Learned

Email as an Interface: Email remains one of the most universal and accessible interfaces for data collection
AI Context Understanding: GPT-4 excels at extracting structured data from unstructured natural language
Real-time User Experience: Immediate feedback loops significantly improve user engagement
Content Moderation: Automated content filtering is essential for user-generated content platforms
Workflow Automation: n8n provides powerful visual workflow design for complex email processing pipelines

🚀 Future Possibilities & Enhancements

Multi-language Support: Expand AI analysis to support multiple languages
Advanced Analytics: Machine learning models for mood prediction and trend analysis
Integration Ecosystem: Connect with fitness trackers, calendar apps, and other wellness tools
Team Features: Shared mood tracking for families or therapy groups
Mobile App: Native mobile app with email composition shortcuts

Screenshots

Mood Trends

Recents Moods

N8N Workflow

✉️ Contact

For questions, suggestions, or collaboration opportunities, please reach out via GitHub Issues or connect with me on GitHub.

Thank you for exploring InboxInsights! 🌟