HonestHomes

knowledgeable friend who happens to have access to every rental listing in Bengaluru

Comment

Inspiration

What it does

How we built it

Challenges we ran into# HonestHomes

🏠 About the Project

Renting in Bengaluru’s competitive market is overwhelming. Renters often face information asymmetry where brokers hide true market rates, red flags stay buried in fine print, and tenants struggle to know whether they are overpaying. We wanted to build a platform that gives renters data-backed confidence to negotiate better deals and avoid rental scams.

HonestHomes is a full-stack AI-powered rental analysis platform that helps users discover trustworthy rental listings with real-time market insights, intelligent search, and AI-driven recommendations.

💡 What it does

HonestHomes combines live rental market data, semantic search, and AI analysis to help renters make informed decisions.

Key Features

  • 🕷️ Crawls 5 major rental platforms daily including 99acres, NoBroker, Housing.com, Magicbricks, and Sulekha.
  • 🔍 Provides hybrid intelligent search using keyword search, semantic understanding with ELSER, and vector similarity.
  • 🤖 Generates AI-powered rental verdicts using Gemini 1.5 Pro to:
    • Flag overpriced listings
    • Detect suspicious patterns and hidden red flags
    • Suggest negotiation strategies based on market trends
    • Provide locality-based market insights
  • 📊 Aggregates rental market statistics by locality, BHK type, and price range.
  • ⚡ Delivers results in real-time through a responsive Next.js dashboard.

Result

Users can confidently compare listings, identify scams, and negotiate better rental deals using live market intelligence.

🛠️ How we built it

Tech Stack

Frontend

  • Next.js 15
  • TypeScript
  • Tailwind CSS

Backend

  • Node.js
  • Express.js

Search & AI

  • Elasticsearch
  • Elastic ELSER
  • Google Gemini 1.5 Pro

Infrastructure & Storage

  • Docker
  • AWS EC2
  • AWS S3
  • Elastic Cloud
  • Vercel

Crawling & Data Collection

  • OpenClaw
  • Firecrawl

Key Components

Rental Crawler

Automated crawler powered by OpenClaw and Firecrawl that scrapes listings from multiple rental websites while handling rate limits and anti-bot protections.

Hybrid Search Engine

Implemented BM25 keyword search, ELSER semantic embeddings, and Reciprocal Rank Fusion (RRF) to provide accurate and context-aware rental search results.

AI Verdict Engine

Gemini-powered analysis system that detects overpriced listings, suspicious broker behavior, outdated information, and provides actionable rental insights.

Data Pipeline

Built a complete ingestion pipeline using AWS S3 for storage, validation, normalization, and Elasticsearch indexing.

Frontend Dashboard

Developed a modern responsive UI with advanced filtering, search panels, locality insights, and AI-generated recommendations.

🔄 Data Flow

Rental Websites
      ↓
Firecrawl Scraper
      ↓
AWS S3
      ↓
Backend Ingestion
      ↓
Elasticsearch
      ↓
Search API
      ↓
Next.js Frontend

🚧 Challenges we ran into

Data Quality Issues

Every rental website used different formats for prices, amenities, and BHK descriptions. We solved this with normalization pipelines, regex transformations, and validation layers.

Duplicate Listings

The same property often appeared across multiple websites. We implemented URL-based deduplication along with content hashing.

Semantic Search Complexity

Traditional keyword search failed for natural rental queries. Integrating ELSER dense vectors with BM25 and RRF significantly improved relevance.

AI Latency

Gemini API responses initially slowed down the application. We implemented caching and asynchronous processing to improve user experience.

Scraper Blocking

Rental websites aggressively block crawlers. We introduced rate limiting, randomized delays, and browser-based scraping using Firecrawl.

Infrastructure Optimization

Continuous scraping became expensive, so we optimized the system with scheduled crawling instead of constant polling.

🏆 Accomplishments that we're proud of

  • ✅ Built a production-ready platform with 6000+ lines of code
  • ✅ Developed a hybrid search system combining 3 intelligent search methods
  • ✅ Created a multi-source crawler supporting 5 rental platforms
  • ✅ Implemented AI-powered red flag detection and market analysis
  • ✅ Successfully deployed the entire stack using Docker, EC2, Elastic Cloud, S3, and Vercel
  • ✅ Added robust logging, monitoring, validation, and error handling
  • ✅ Built locality-based rental aggregation and analytics
  • ✅ Documented the complete architecture, deployment flow, and APIs

📚 What we learned

  • Building reliable data pipelines is harder than expected.
  • Semantic search performs significantly better than basic keyword search for real-world rental queries.
  • Browser-based scraping is more effective than traditional HTML parsing.
  • LLM cost optimization and caching are essential for scalable AI applications.
  • Infrastructure, deployment, and monitoring take as much effort as feature development.
  • Early architecture decisions strongly impact scalability and maintainability.

🚀 What's next for HonestHomes

Multi-city Expansion

Expand beyond Bengaluru into cities like Mumbai, Delhi, and Pune.

Rental Price Prediction

Build machine learning models that estimate fair rental pricing based on amenities, locality, and historical trends.

Community Reviews

Allow renters to leave verified reviews and broker ratings.

Mobile Application

Launch React Native apps for Android and iOS with push notifications for new listings and price drops.

Broker Accountability

Track broker behavior, complaint patterns, and suspicious activities.

Smart Alerts

Notify users when listings matching their preferences become available.

Payment & Verification

Enable secure deposits and direct booking workflows for verified listings.

Community Insights

Aggregate renter experiences and safety insights locality-wise.

🧰 Built With

Languages

  • JavaScript
  • TypeScript

Frontend

  • Next.js 15
  • Tailwind CSS

Backend

  • Node.js
  • Express.js

Search & AI

  • Elasticsearch
  • Elastic ELSER
  • Google Gemini 1.5 Pro

Cloud & Infrastructure

  • AWS EC2
  • AWS S3
  • Elastic Cloud
  • Docker
  • Vercel

Tools & Services

  • Firecrawl
  • OpenClaw
  • GitHub

🔗 Try it Out

Accomplishments that we're proud of

What we learned

What's next for HonestHomes

Built With

Share this project:

Updates