Python, Git, and API keys
You already installed Cursor or Claude Code. Use it for Python setup too - do not hunt for install commands yourself unless the agent gets stuck. You still confirm a few things yourself (version output, GitHub account, API key) because those are account facts the agent cannot complete for you.
Step 1: Python setup (agent builds, you verify)
I am setting up for an AI engineering bootcamp. Check whether Python 3.11+ is installed on my machine (run the right version command for my OS). If it is missing or too old, tell me the exact install steps for my OS and run what you can. Then create a virtual environment in this folder, give me the exact activate command, add a .gitignore that excludes .venv and .env, and confirm pip works inside the venv.
Run whatever install steps it gives you (it may ask you to approve shell commands). When it is done, you should see Python 3.11+ and an activated venv.
This course targets Python 3.11 or newer (3.12 is fine). After your agent runs, double-check in the terminal:
python3 --version
python --version
py --version
You want output like `Python 3.11.9` or `Python 3.12.x`. Anything 3.10 or older - paste the output back to your agent and ask it to fix the install.
Tip
If the agent cannot install Python
Some installs need your password or a GUI installer the agent cannot click through. Fallback: Mac - python.org/downloads or `brew install python@3.12`. Windows - python.org installer, check Add python.exe to PATH. Linux - `sudo apt install python3.12 python3.12-venv`. Then tell your agent: `Python is installed now - create the venv and .gitignore in this folder.`
Step 2: GitHub account (required before Session 1)
Session 1 deploys from a GitHub repo fork. Create a free GitHub account at github.com if you do not have one - use the email you check regularly.
Critical
New to GitHub? Do this course first
Work through Introduction to GitHub on GitHub Learn (free, ~30 minutes). It covers creating an account, repos, commits, and pull requests - enough to fork the Session 1 starter repo and follow along. Your coding agent can help after you have the basics; it cannot sign up for GitHub on your behalf.
Optional: ask your agent to check if git is installed. On Windows, Claude Code may need Git for Windows if you hit bash errors.
Check if git is installed on my machine and tell me the version.
Step 3: OpenAI API key (required for Session 1)
Critical
Required for Session 1
Session 1 calls OpenAI from FastAPI. You need an API key before the live session. Ask your agent to create `.env` and `.env.example` - you paste the key locally, never into chat.
OpenAI API key
Required for Session 1Sign in at platform.openai.com. Go to API keys and create a key. Store it in `.env` as `OPENAI_API_KEY=sk-...` - never commit it to Git. If you hit rate limits during builds, add billing credit under Billing on the same site.
Watch out
Secrets hygiene
- Pasting live API keys into Maven, WhatsApp, or screenshots.
- Committing `.env` to GitHub - only commit `.env.example` with placeholder names.
Anthropic keys come later - you will need one before Session 3 for agent weeks. Good to have ready if you already use Claude Code, but not required before Session 1.
Step 4: FastAPI hello-world (agent builds, you run it)
With Python, venv, and `.env` in place, paste the prompt below. Then install dependencies and run the server. The POST /ask smoke test is optional - we build that in Session 1.
Create a minimal FastAPI hello-world app in main.py with GET /health returning {"status": "ok"}, add requirements.txt with fastapi and uvicorn, and tell me how to run it.pip install -r requirements.txt
uvicorn main:app --reload
Open http://127.0.0.1:8000/health in your browser. You should see JSON like `{"status":"ok"}`. This confirms Python and pip work.
Ready when
Ready when
- `python3 --version` (or `python --version`) shows 3.11+
- A virtual environment is created and activated in your project folder
- GitHub account created (Introduction to GitHub course done or skipped if you already use GitHub)
- OpenAI API key is created and lives in `.env` (not in chat)
- GET /health returns JSON locally (good to have - we can build this in Session 1 if needed)
Done when
- Python 3.11+ confirmed with a version command
- Virtual environment created and activated
- GitHub account created
- OpenAI API key in `.env`
- Local FastAPI /health endpoint responds (optional before Session 1)