🧠 LangChain LLM Apps: From Simple Chains to Stateful Chatbots and RAG

This repository is a comprehensive, hands-on walkthrough for building production-grade LLM applications using LangChain, Groq, OpenAI, and ChromaDB. It includes everything from a basic LCEL chain to a full-fledged chatbot with memory, and culminates with a Retrieval-Augmented Generation (RAG) pipeline using vector stores and retrievers.

📁 Project Structure

This repo is organized into three main notebooks (or modules):

1. 01_simple_llm_app.ipynb - A minimalist LangChain Expression Language (LCEL) pipeline to translate text into another language.
2. 02_chatbots.ipynb - A conversational chatbot that remembers previous user inputs using RunnableWithMessageHistory.
3. 03_vectorretriever.ipynb - Introduction to document embeddings, vector stores, retrievers, and a basic RAG implementation using LangChain components.

🚀 What You’ll Learn

🧱 Core Concepts

• How to make LLM calls via LangChain with OpenAI or Groq (Gemma/LLaMA3).
• Crafting prompts with ChatPromptTemplate.
• Chaining components using LangChain Expression Language (LCEL).
• Parsing outputs with StrOutputParser.

💬 Stateful Chatbots

• Implementing chat history with RunnableWithMessageHistory.
• Managing session-based interactions.
• Trimming messages to fit LLM context windows.
• Injecting system-level instructions dynamically.

📚 Vector Search + RAG

• Creating semantic search pipelines with HuggingFace Embeddings.
• Using ChromaDB to store and retrieve relevant chunks.
• Wrapping retrievers into LCEL-compatible chains.
• Building basic Retrieval-Augmented Generation (RAG) systems.

🛠️ Tech Stack

• LangChain (core + community integrations)
• Groq (Gemma 2B / LLaMA 3 via langchain_groq)
• OpenAI GPT-4 / 3.5
• ChromaDB for local vector storage
• HuggingFace Embeddings (all-MiniLM-L6-v2)
• LangSmith (optional) – for tracing/debugging
• Python 3.10+

📦 Setup Instructions

# 1. Clone the repository
git clone https://github.com/FarazF19/Langchain-Chatbots_RAG.git
cd Langchain-Chatbots_RAG


# 2. Install dependencies
pip install -r requirements.txt

# 3. Add your API keys
touch .env
# Add your keys
OPENAI_API_KEY=your_openai_key
GROQ_API_KEY=your_groq_key
HF_TOKEN=your_huggingface_token

🧪 Example Usage

Simple Translator Chain

chain.invoke({"language": "French", "text": "Hello, how are you?"})

Chatbot with Memory

with_message_history.invoke(
    [HumanMessage(content="What is my name?")],
    config={"configurable": {"session_id": "chat3"}},
)

Retrieval-Augmented Response

rag_chain.invoke("tell me about dogs")

🧠 Key LangChain Concepts Highlighted

| Feature | Description | | ---------------------------- | --------------------------------------------------------- | ---------------------------------------------------- | | ChatPromptTemplate | Modular prompt templates for structured messaging | | StrOutputParser | Clean extraction of string responses from raw LLM outputs | | RunnableWithMessageHistory | Maintains chat state across multiple LLM calls | | Chroma.from_documents | Create vector store from document chunks with embeddings | | as_retriever() | Convert vector store into retriever for use in RAG | | LCEL ( | operator) | Expressive pipeline chaining of LangChain components |

---

📈 Future Enhancements

• Add LangGraph and Crew AI multi-agent support
• LangSmith Tracing & Observability Integration
• RAG Fusion (hybrid search techniques)
• Frontend deployment using LangServe / FastAPI

🙌 Credits

• LangChain
• Chroma
• Groq
• OpenAI
• HuggingFace

Tutor

Kirish Naik

👨‍💻 Maintainer

Muhammad Faraz
AI Full Stack Developer
LinkedIn •