An intelligent Retrieval-Augmented Generation (RAG) agent powered by Groq AI and local embeddings, designed to process and understand your custom text data efficiently — all without relying on expensive external APIs.
This project demonstrates how to build a lightweight, privacy-friendly AI agent that can retrieve contextually relevant information from your own data and generate accurate, natural-language responses using Groq’s blazing-fast inference engine.
- 🔍 Retrieval-Augmented Generation (RAG) – Combines information retrieval with language generation for context-aware responses.
- ⚡ Groq API Integration – Utilizes Groq’s high-speed inference models (e.g.,
llama-3.1-8b-instant). - 🧩 Local Text Embeddings – Uses
sentence-transformersto embed and index text data efficiently with ChromaDB. - 🔒 Privacy-Friendly – Keeps your data local; no sensitive data leaves your machine.
- 🧠 Fully Modular – Easy to extend with your own datasets or models.
| Component | Description |
|---|---|
| Python 3.11+ | Core programming language |
| Groq API | LLM inference engine |
| LangChain | Framework for RAG and data orchestration |
| ChromaDB | Vector database for document retrieval |
| Sentence Transformers | Local embedding model (all-MiniLM-L6-v2) |
| dotenv | Environment variable management |
git clone https://github.com/harshithreddyv9/rag_agent_groq_v2.git
cd rag_agent_groq_v2