An AI-powered assistant that helps you find and analyze real estate properties in Argentina using ZonaProp listings. It is not a real real estate agent.

• Natural Language Search: Simply describe what you're looking for, and the assistant will generate the appropriate ZonaProp search URL
• Automated Scraping: Collects property listings from ZonaProp based on your search criteria
• Intelligent Analysis: Analyzes property data to provide insights on pricing, location, features, and value
• Personalized Recommendations: Suggests 3-5 properties that best match your requirements with pros and cons for each

This application uses a state graph architecture with the following components:

1. Query Data Agent: Converts natural language queries into structured data for targeted web search
2. Scraper: Collects property data from ZonaProp using pagination
3. Analysis Agent: Evaluates properties based on user criteria
4. Recommendation Agent: Provides personalized property recommendations

• Python 3.13+
• Required Python packages (install via pip install -r requirements.txt or uv add -r requirements.txt)
• Docker (Optional, for web gui)
• AWS Serverless Application Model Command Line Interface (AWS SAM CLI) (Optional, for web gui)
• Ollama gemma3:12b and apropiate hardware (Optional)

1. Clone this repository:

   git clone https://github.com/okason97/Not-Real-Estate-Agent.git
   cd zonaprop-real-estate-agent

2. Install dependencies:

   uv add -r requirements.txt

1. Set up environment variables (If you want to use an API):

   # Create .env file
   touch .env
   # Add necessary environment variables

2. Or install and run Ollama with the gemma3:12b model locally:

   # Install Ollama: https://ollama.com/
   # Then pull the model
   ollama pull gemma3:12b
   ollama run gemma3:12b

3. Run the notebook:

   code agent.ipynb

1. Build the Docker image and SAM (Serverless Application Model):

   docker build --no-cache --provenance=false --platform linux/amd64 -t nrea .
   sam build --no-cached

2. Open Ollama server (Optional):

   ollama serve

3. Start a local API Gateway at http://127.0.0.1:3000/ using SAM (Serverless Application Model):

   sam local start-api

4. Run the web GUI:

   streamlit run app.py

Enter a message: I'm looking for a 2-bedroom apartment to rent in La Plata, with a garage, preferably less than 10 years old.

[The system generates a ZonaProp URL, scrapes listings, analyzes them, and provides recommendations]

• agent.ipynb: Notebook to try the agents
• frontend.py: Serves the interactive frontend
• app/: Directory containing the agents and the API code
  • main.py: Contains the FastAPI behaviour
  • src/: Directory containing the source code
    • scraper.py: Contains the Scrapper classes for web scraping
    • graph.py: Contains the graph building logic
    • models.py: Loads the LLM model
    • nodes.py: Contains the agentic nodes of the graph
    • state.py: Contains the internal state of the graph

You can customize the application by:

1. Modifying the prompt templates in each agent
2. Extending the scraper to collect information from other sites by subclassing Scrapper
3. Extending the scraper to collect additional property details
4. Adjusting the analysis criteria based on your preferences

• Add a plan-and-execute architecture for the agents
• Implement historical price tracking and other useful information for the analysis agent
• Add visualization tools for property comparisons
• Support for additional property listing websites
• Add an image processing gent to include visual information

This project is licensed under the MIT License - see the LICENSE file for details.

This tool is for educational purposes only. Please respect ZonaProp's terms of service and implement appropriate rate limiting in production environments.