⚡ VoltIQ - Energy Intelligence RAG

An AI-powered Hybrid RAG (Retrieval-Augmented Generation) system for answering energy-related questions using both unstructured documents and structured energy datasets.

📸 Interface

The system combines document retrieval, tabular analytics, vector search, query routing, and an LLM to provide data-driven answers through a custom, fast, and responsive web interface.

---

🚀 Features

• Custom Web Interface: Modern, responsive UI built with pure HTML/CSS/JS served via FastAPI.
• Low-Memory Footprint: Uses fastembed for blazing-fast, local embedding generation without the massive PyTorch overhead, making it perfect for free-tier deployments.
• Intelligent Query Routing: Dynamically routes questions between document retrieval and tabular analytics.
• Hybrid LLM Backend: Primary reasoning powered by Groq, with an automatic failover to OpenRouter.
• Persistent Vector Search: Built-in ChromaDB for fast and reliable document retrieval.
• Energy Analytics: Understands energy conservation, consumption analytics, weather impacts, and tariff comparisons.

---

📊 Data Sources & Capabilities

VoltIQ is grounded in specific, real-world data and documents. It can instantly answer questions based on:

1. London Household Energy Data (Tabular): Analytics on average daily energy consumption across thousands of households, including the impacts of weather, seasons, and bank holidays.
2. Energy Tariffs: Comparisons between Standard and Time-of-Use tariffs.
3. Conservation Documents (PDF/TXT): Unstructured knowledge base containing expert recommendations, tips, and strategies for industrial and residential energy conservation.

Whenever you ask a question, the AI intelligently decides whether to run mathematical queries on the tabular data or perform a vector search across the documents to find your answer!

---

🛠️ Tech Stack

• Backend: FastAPI (Python)
• Frontend: HTML5, CSS3, Vanilla JavaScript
• Vector Database: ChromaDB
• Embedding Engine: FastEmbed (BAAI/bge-small-en-v1.5)
• LLM Providers: Groq API, OpenRouter API
• Data Processing: Pandas

---

🧠 System Architecture

                     User Question (Web UI)
                               │
                               ▼
                        FastAPI Server
                               │
                               ▼
                        Query Router
                               │
                ┌──────────────┴──────────────┐
                ▼                             ▼
         Document Queries                Data Queries
                │                             │
                ▼                             ▼
  FastEmbed (bge-small-en-v1.5)        Pandas Analytics
                │                             │
                ▼                             ▼
         Chroma Vector DB                Data Processing
                │                             │
                └──────────────┬──────────────┘
                               ▼
                    Combined Context / Data
                               │
                               ▼
                       LLM Generation
                (Groq -> Fallback OpenRouter)
                               │
                               ▼
                        Generated Answer
                               │
                               ▼
                          VoltIQ UI

---