This project demonstrates how to build an intelligent chatbot with memory capabilities using LangChain, powered by the Groq LLM (Gemma2-9b-It). It includes multilingual support, message trimming for long conversations, and conversation history via session management.
1-chatbots.ipynb– Main notebook containing all the logic for:- Connecting to Groq API
- Maintaining chat history
- Using prompt templates
- Supporting multiple languages
- Trimming chat history to fit within token limits
git clone https://github.com/your-username/langchain-groq-chatbot.git
cd langchain-groq-chatbot
-
Create a virtual environment
python -m venv venv
venv\Scripts\activate
source venv/bin/activate
3.Install dependencies
pip install -r requirements.txt
4.Set up your .env file
Create a .env file in the root directory and add your Groq API key:
GROQ_API_KEY=your_groq_api_key_here
🚀 Running the Notebook To run the chatbot:
jupyter notebook Open 1-chatbots.ipynb and run cells step-by-step.
💬 Features ✅ Memory across sessions using RunnableWithMessageHistory
✅ Dynamic prompt templating with system messages
✅ Language-aware responses (e.g., English, Hindi)
✅ Token-aware message trimming via trim_messages
✅ Powered by Groq’s Gemma2-9b-It model
🌐 Example
HumanMessage(content="Hi, my name is Arjun")
HumanMessage(content="What's my name?") The assistant will remember that your name is Arjun if you're using the same session.
You can also specify response language:
{ "messages": [HumanMessage(content="Hi")], "language": "Hindi" } 📦 Dependencies Make sure these packages are included in requirements.txt:
langchain langchain-core langchain-groq langchain-community python-dotenv tqdm 🧠 Notes on History Management To avoid overflowing the LLM's context window:
Message history is trimmed using trim_messages
Only the last few tokens are retained based on a limit (max_tokens=45)
You can control inclusion of system messages and trimming strategy