Day 1 - Environment Setup Overview: UV, API Keys, Costs and GitHub Repo

Week 1 · Day 1 · Lecture 4. Setting up your environment is the rite of passage before any of the interesting work. Most of the time it goes smoothly. Occasionally it does not — and this lecture is about being equipped for both.

The three things this lecture sets up

1. uv — a Python package manager that is fast and bulletproof, and does most of the environment work for you.
2. API keys — a .env file holding the credentials that connect your code to models like OpenAI.
3. The GitHub repo — the labs, the guides and the troubleshooting notes, all kept up to date.

Once it is done, it is done. You will not repeat this.

What the course will cost you

This course calls real models running in the cloud, and those calls have a price. The honest numbers:

There is no need to spend a penny on this course. Your options:

• OpenAI — the default; keeps you in step with the lectures.
• Anthropic Claude — a strong alternative.
• Google Gemini — has a free tier.
• OpenRouter — one key, many free and cheap models.
• DeepSeek — very cheap.
• Ollama — models running locally on your own machine, free.

All of it is documented in Guide 9 (guides/09_ai_apis_and_ollama.ipynb).

If you go the free route, expect to work a little harder: smaller steps, more experimentation, more iteration to reach good output. That is normal, and it is completely workable.

Practical steps

1. Get the repo

Clone the course GitHub repository. It contains the labs, the guides/ folder, the setup/ instructions and troubleshooting notes. The labs are updated regularly — work from a current copy, not a stale download.

2. Follow the setup guide for your OS

• Windows → setup/SETUP-PC.md
• macOS → setup/SETUP-mac.md
• Linux → setup/SETUP-linux.md

3. Install uv and sync the environment

uv reads the project's pyproject.toml and builds the virtual environment for you. One command does the whole job:

uv sync

4. Create your .env file

In the project root, create a file called .env holding your keys:

OPENAI_API_KEY=sk-proj-...

Spell the variable name exactly. The single most common cause of a broken setup is writing OPEN_API_KEY instead of OPENAI_API_KEY. It costs people hours. Copy it, do not type it.

5. Choose your model path

Decide now whether you are running on OpenAI, on a free tier, or on local models, and set the matching keys. Guide 9 has the exact variable names for each provider.

Keep these front of mind

1. The resources are extensive — use them

There is a course website with links and extra videos, a GitHub repo with the guides and troubleshooting, and the labs themselves. The instructions written inside each lab notebook carry the most current information — read them, do not skim past to the code.

2. Stay patient when something breaks

There will be irritating moments in the coming weeks. Problem-solving is where the real learning happens, and working through an obstacle builds a toolkit you keep. Try to enjoy it rather than fight it.

3. Check the simple things first

When something seems deeply strange, there is almost always a simple explanation waiting at the end of it — a typo in a variable name, a key in the wrong file, a stale copy of a lab. Check those before assuming something profound is wrong.

You are done when

• uv sync completes without errors.
• A .env file exists in the project root with correctly spelled keys.
• You can open and run a lab notebook.
• You know which provider you are using and roughly what it will cost you.

Resource links

• Course resources page — links and supplementary videos
• uv documentation (Astral)
• guides/09_ai_apis_and_ollama.ipynb — Guide 9: free and cheap model options, provider by provider
• setup/SETUP-PC.md · setup/SETUP-mac.md · setup/SETUP-linux.md
• OpenAI usage dashboard — watch your spend
• OpenRouter · Ollama · Google AI Studio (Gemini)
• Instructor Q&A avatar — answers to the most common course questions

Next

Lectures 1.5 and 1.6 walk through the Windows and Mac setups step by step; 1.7 sets up your OpenAI key and .env file for the first lab.