Chapter 03 CLI Chatbot

Minimal single-file CLI chatbot for Chapter 3 of LLM Engineering in Action.

The default configuration targets OpenRouter through the OpenAI Python SDK’s OpenAI-compatible interface.

Features

• OpenAI chat-completions style message array
• OpenRouter as the default OpenAI-compatible provider
• Single-file Python implementation
• Streaming responses with stream=True
• Token usage tracking from the final stream metadata chunk
• Basic per-session cost estimation
• CLI commands for clearing history and showing stats
• OPENAI_BASE_URL=https://openrouter.ai/api/v1
• OPENAI_MODEL=openai/gpt-5-nano

Setup

python -m venv .venv
.venv\Scripts\activate
pip install -r requirements.txt
copy .env.example .env

Then set OPENAI_API_KEY in .env.

OpenAI-Compatible Providers

This project uses the official openai Python SDK, but the actual provider can be any service that supports the OpenAI chat completions request format. OpenRouter is one such provider, so the same code works by changing the base URL and model name rather than rewriting the app.

In this sample, the defaults are:

OPENAI_BASE_URL=https://openrouter.ai/api/v1
OPENAI_MODEL=openai/gpt-5-nano

That means:

• The SDK still sends client.chat.completions.create(...)
• Requests go to OpenRouter instead of OpenAI’s own endpoint
• The model name uses OpenRouter’s routing format

Run

python chatbot.py

Commands

• /help shows commands
• /clear resets chat history
• /stats prints token and cost totals
• /exit quits the app

Project Structure

chapter-03-chatbot/
├── chatbot.py
├── requirements.txt
├── .env.example
└── .gitignore