README.md

AURA AI Service

This service provides the intelligent backbone for AURA, handling document processing, Knowledge Base (RAG) search, and LLM reasoning.

🚀 Overview

The AI Service is built with FastAPI and uses a RAG (Retrieval-Augmented Generation) pattern to grounded AURA's responses in specific documents.

🛠 Tech Stack

• Framework: FastAPI (Python 3.10+)
• Vector Search: Qdrant (local or cloud)
• Embeddings: Sentence-Transformers
• ORMs/Tools: Pydantic, SQLAlchemy

📋 Capabilities

• Document Ingestion: Upload .txt, .pdf, and .pptx documents to AURA's brain.
• RAG Search: Semantic search over uploaded documents to provide context to the Voice Agent.
• API Endpoints:
  • GET /api/v1/rag/search: Search the knowledge base.
  • POST /api/v1/rag/upload: Ingest new documents.

⚙️ Setup

1. Navigate to directory: cd ai-service
2. Setup environment: python -m venv venv
3. Activate environment:
   • Windows: venv\Scripts\activate
   • Unix: source venv/bin/activate
4. Install dependencies: pip install -r requirements.txt
5. Configure .env:
   • OPENAPI_KEY: Your LLM provider key.
   • QDRANT_URL: URL to your vector database.
6. Run: uvicorn app.main:app --reload