Real Estate Agent Tools

A platform for independent real estate agents and small brokerages that streamlines day-to-day workflows and adds data intelligence that larger platforms don't provide. Built for the Indiana/Kentucky market.

What It Does

• Property management — add and manage listings and buyer-side representations with full parcel data, county boundaries, and map overlays
• AI property chat — ask questions about a property using uploaded documents and property details as context
• Net sheet — seller net proceeds calculator with real property tax proration (Indiana DLGF integration)
• Document management — upload, view, and delete property documents
• Crime heatmap — raster tile overlay for violent and property crime (Louisville metro)
• Social post generation — AI-generated listing copy from property details and documents; multi-platform posting support in progress
• Multi-tenant auth — brokerage and agent accounts with JWT authentication

See docs/ROADMAP.md for what's planned next.

---

Tech Stack

Backend

• Python 3.11, FastAPI, SQLModel (SQLite)
• poetry for package management
• OpenAI API (chat and embeddings)
• Milvus (document vector search)
• AWS S3 (document storage)
• ArcGIS IGIS (parcel data)
• Indiana DLGF (property tax data)

Frontend

• React 18, Tailwind CSS
• Mapbox GL / react-map-gl
• React Router

Pipeline

• Separate Python/Prefect service for data ingestion (crime heatmap, property tax)

---

Running Locally

Prerequisites

• Python 3.11+
• Node 18+
• poetry (pip install poetry)
• Docker and Docker Compose (for Milvus)

1. Start supporting services

docker-compose up -d

This starts Milvus (vector database) and its dependencies (etcd, MinIO).

2. Backend

cd backend
cp .env.example .env   # fill in your API keys
poetry install
poetry run uvicorn backend.main:app --reload

The backend runs on http://localhost:3000.

Required environment variables (see .env.example for full list):

OPENAI_API_KEY=
RENTCAST_API_KEY=
GOOGLE_MAPS_API_KEY=
MILVUS_URI=http://localhost:19530
S3_BUCKET=
S3_ACCESS_KEY=
S3_SECRET_KEY=
JWT_SECRET_KEY=

3. Frontend

cd frontend
npm install
npm start

The frontend runs on http://localhost:3001 and expects the backend at http://localhost:3000.

4. Pipeline (optional)

The pipeline ingests crime and property tax data. Only needed if you are working on data pipeline features.

cd pipeline
poetry install
poetry run python -m pipeline.flows.<flow_name>

---

Running Tests

# Backend
cd backend && poetry run pytest

# Frontend
cd frontend && npm test

# Pipeline
cd pipeline && poetry run pytest

---

Project Structure

backend/        FastAPI backend, SQLModel models, services, repositories
frontend/       React frontend
pipeline/       Data ingestion pipeline (crime heatmap, property tax)
docs/           Product documentation and roadmap

---

Database

The database is a SQLite file at backend/real_estate.db. It is created automatically on first run.

Schema migrations are currently handled by a temporary script while the schema foundation is being established. This will be replaced with Alembic once the schema stabilises.

cd backend

# Apply schema changes (temporary — will be replaced by Alembic)
poetry run python scripts/migrate_db.py

# Migrate data from legacy schema (one-time, if upgrading from an older version)
poetry run python scripts/migrate_property_info.py