Skip to content

ImmoSearch

We provide you with your own real estate agent that helps you find the appartment of your dreams.

Code Video Cerebral Valley

Core trick

Queries French DVF property transaction data and ranks listings by price against historical medians and commute time.

Tags

financeproductivity

Description

We provide you with your own real estate agent that helps you find the appartment of your dreams. Whether for rent or buying, our agent searches the market, ranks listings based on relevance, price compared to historical market price and distance from work, all the steps that you would be doing manually during an appartment hunt. It finally returns a shortlisted table with all the good deals it finds and a map visualizing the location of the appartments and nearby places
imadezze/ImmoSearch★ 00 forksPythonCreated Sep 13, 2025 · last push Sep 14, 2025

README excerpt

# ImmoSearch

MCP server exposing DVF (Demandes de Valeurs Foncières) analysis tools for French real estate data. This project includes both a powerful MCP server for property analysis and a beautiful conversational UI for apartment viewing reservations.

## Features

- 🏠 **Real Estate Analysis**: DVF data analysis with property search and filtering
- 🚗 **Travel Time Calculations**: Automatic commute time calculations to workplace
- 📍 **Location Services**: Reverse geocoding and address lookup
- 💬 **Conversational UI**: Elegant apartment viewing assistant interface
- 📊 **Metrics Integration**: W&B integration for analytics and tracing
- 🌐 **Multiple Interfaces**: Both MCP server and FastAPI web interface

## Project Structure

- `mcp_servers/immosearch_server.py` - Main MCP server for real estate analysis
- `conversationImmoSearch/` - FastAPI web interface with conversational UI
- `voiceassistant/` - LiveKit voice assistant integration
- `scripts/` - Utility scripts for data processing and integrations

## Running the MCP Server

In order to run your MCP server locally, use the following command:

```bash
python -m mcp_servers.immosearch_server
```

The server will automatically start in **streamable HTTP** mode

## Running the Conversational UI

To run the FastAPI web interface with the apartment viewing assistant:

```bash
cd conversationImmoSearch
python app.py
```

Then visit `http://localhost:7860` to access the conversational interface.

## Deployment Commands

For depl

Team