generativeai-dev

This repository contains a collection of examples for building conversational AI and LLM-powered applications using Chainlit.

🚀 Overview

This project provides several sample scripts demonstrating how to integrate Chainlit with various powerful tools from the LLM ecosystem. The examples include:

• A basic chat interface with an LLM.
• An advanced agent-like application using LangGraph to create stateful, multi-step chains.
• An integration with Ollama to run and interact with open-source LLMs locally on your machine.

✨ Features

• Interactive chat interface powered by Chainlit.
• Examples of building stateful agents with LangGraph.
• Demonstration of using local LLMs via Ollama.

📋 Requirements

• Python 3.8+
• Chainlit
• LangChain (langchain, langchain-openai, etc.)
• LangGraph (langgraph)
• Ollama (ollama)

For local model usage, you also need to have the Ollama application installed and the desired models pulled (e.g., ollama pull llama3).

📦 Installation

This project uses uv for fast Python package management. If you don't have it installed, follow the instructions on their website.

1. Clone the repository:

   git clone https://github.com/tanaka-r/generativeai-dev.git
   cd generativeai-dev

2. Create a virtual environment and install the dependencies from pyproject.toml using uv:

   # Create and activate a virtual environment
   uv venv
   source .venv/bin/activate
   
   # Install dependencies defined in pyproject.toml
   uv pip install -e .

▶️ Running the Examples

To run a specific example, use the chainlit run command followed by the path to the script. The -w flag enables auto-reloading when you make changes to the code.

# Example for a generic app
chainlit run app/chainlit_app.py -w

(Replace app.py with the name of your main Python script if it's different.)

Then, open your web browser and navigate to http://localhost:8000.