You can access the live application here: https://greenvisor.vercel.app/
Version: 0.0.1
Last Updated: May 10, 2025
For detailed documentation on specific parts of the project, please refer to:
Greenvisor is an AI-powered energy consultation application designed to provide advice on home renovations for energy efficiency, with a focus on the German context. It leverages Large Language Models (LLMs) to answer user questions, considering German laws, government subsidies, and potential cost savings. The application can interact with a Neo4j graph database for retrieving contextual information and can export data to PDF format.
- AI-Powered Energy Consultation: Provides tailored advice on home energy renovations.
- German Context-Specific: Considers German laws (e.g., GEG §72), subsidies, and common renovation practices.
- Structured JSON Output: Delivers renovation suggestions in a predefined JSON format, detailing measures, costs, savings, CO2 impact, subsidies, and legal obligations.
- Multi-Language Support: Capable of understanding prompts and delivering answers in different languages (e.g., English, German).
- Graph Database Integration: Utilizes a Neo4j graph database to fetch relevant data and context for answering user queries.
- PDF Export Functionality: Allows for the export of data (presumably consultation results or reports) into PDF files.
- Modular Design: Built with FastAPI for the API layer and Langchain for LLM interactions.
- Enhanced Reports: Add customizable visualization and detailed PDF export options.
- Automated Subsidy Updates: Automatically update subsidy information as government policies change, relevant to current user data and chat history.
- Human Expert Feedback System: Integrate a system to receive feedback and impovemnet of the outputs from an Human.