Build a Language‑Learning Agent with OpenAI, LangGraph, Ollama & MCP

A hands-on project demonstrating how to build an AI-powered language-learning agent by combining:

• OpenAI – cloud LLM for high-quality language generation and evaluation
• LangGraph – stateful agent orchestration with graph-based control flow
• Ollama – local LLM inference for offline/private usage
• MCP (Model Context Protocol) – standardized tool and context integration

Tech Stack

Component	Purpose
OpenAI API	Cloud-based LLM (GPT-4o)
LangGraph	Agent graph & state management
Ollama	Local model serving
MCP	Tool protocol & context injection
Python 3.13	Runtime

Getting Started

# Create and activate virtual environment
python -m venv .venv
source .venv/bin/activate

# Install dependencies
pip install -r requirements.txt

# Run the agent
python main.py

Environment Variables

OPENAI_API_KEY=your_openai_api_key