📚 LangChain RAG Bot

A lightweight Retrieval-Augmented Generation (RAG) bot built on LangChain docs. This assistant can answer technical questions such as:

“How do I define a custom agent in LangChain?”

“What’s the role of memory in LangGraph?”

It uses:

HuggingFace embeddings for semantic understanding

ChromaDB as a vector store with cosine similarity

Groq LLM (LLaMA-3) for fast, accurate responses

✨ Purpose

The project demonstrates how to build a domain-specific knowledge assistant using only LangChain’s official documentation (20–50 curated pages). This keeps the knowledge base focused, efficient, and evaluator-friendly.

⚙️ Methodology

Load Docs → Fetch LangChain documentation pages.

📌 Project Scope

• Domain: LangChain official documentation only
• Knowledge Base: 20–50 curated pages
• Retrieval: ChromaDB embeddings with cosine similarity
• Model: Groq LLaMA-3 for answer generation

Clean & Chunk → Split docs into overlapping chunks for context.

Embed Text → Use HuggingFace all-MiniLM-L6-v2 for embeddings.

Store in Vector DB → Save embeddings in ChromaDB (cosine similarity).

Retrieve Relevant Chunks → Query vector DB for relevant sections.

Generate Answer → Use Groq’s LLaMA-3 model with RAG context.

📋 Prerequisites

Python 3.9+

Virtual environment (venv) recommended

Groq API key (free)

🛠 Installation

1️⃣ Clone the repo

git clone https://github.com/your-username/langchain-rag-bot.git cd langchain-rag-bot

2️⃣ Create & activate virtual environment

python3 -m venv venv source venv/bin/activate # macOS/Linux venv\Scripts\activate # Windows

3️⃣ Install dependencies

pip install -r requirements.txt

🔑 Environment Setup

create a .env file in the project root:

GROQ_API_KEY=your_groq_api_key_here

🚀 Usage

Run the bot from the project root:

python main.py

You can then ask questions like:

❓ What are LangChain agents? ❓ How do retrievers work? ❓ How do I install LangChain?

The bot will:

Retrieve the most relevant chunks from docs

Generate an accurate answer with sources

🧩 Project Structure

langchain_rag_bot/ │── config.py # Configuration & constants │── data_loader.py # Load & clean LangChain docs │── vectorstore.py # Store/retrieve from ChromaDB │── llm.py # Groq LLM integration │── main.py # Entry point │── requirements.txt # Dependencies │── .env.example # Example env file

📝 Example Query

Q: How do I define a custom agent in LangChain?

AI Answer:
Custom agents are built by ... (generated answer here)

Sources:

• concepts/agents/
• concepts/tools/

⚡ Tech Stack

LangChain ChromaDB HuggingFace Sentence-Transformers Groq LLaMA-3

📌 Project Scope

• Domain: LangChain official documentation only
• Knowledge Base: 20–50 curated pages
• Retrieval: ChromaDB embeddings with cosine similarity
• Model: Groq LLaMA-3 for answer generation

📌 Notes for Evaluators

Knowledge base limited to 8 LangChain doc pages (within 20–50 page scope).

Uses cosine similarity for retrieval.

No paid services — HuggingFace (free embeddings) + Groq (free API key).

📜 License

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