Agentic AI course: get your laptop ready
What to install before the first session of the Hackerspace Valencia agentic AI course: git, uv, opencode and the practice repo.
In the agentic AI course for Hackerspace Valencia members we work with agents on our own laptops from minute one. So we don’t lose the first half hour installing things, here is what to bring ready. It takes about 15 minutes.
The course itself is in Spanish, but the slides, the speaker notes and the practice repo are also available in English, and the setup is the same anywhere. If you get stuck, come 30 minutes before the first session and we’ll sort it out together.
What you’ll install
| What | Why |
|---|---|
| git | Saving versions and undoing what the agent does |
| uv | Python and its dependencies. It installs Python for you |
| opencode | The coding agent we use in the course |
| The practice repo | The project we do the exercises on, in English |
You get the OpenRouter key for the models during the session. No payment and no account needed.
Windows: WSL first
On Windows the whole course runs inside WSL, a Linux (Ubuntu) built into Windows. Open PowerShell as administrator and run:
wsl --install
Restart, open the Ubuntu app and create your user. From here on, every step goes in that Ubuntu terminal, not in PowerShell.
1. Git
Check whether you already have it:
git --version
If it doesn’t answer:
- Mac:
xcode-select --install - Ubuntu or WSL:
sudo apt update && sudo apt install -y git
2. uv
uv manages Python and the project dependencies. You don’t need to install Python separately: uv downloads the version the project asks for. It also saves you the certificate error (CERTIFICATE_VERIFY_FAILED) that the macOS Python sometimes throws.
curl -LsSf https://astral.sh/uv/install.sh | sh
On a Mac with Homebrew, brew install uv works too. The other options are in the uv docs.
3. opencode
opencode is an open coding agent that runs in the terminal and works with any model. What we learn with it applies just as well to Claude Code, Codex or Cursor.
curl -fsSL https://opencode.ai/install | bash
On a Mac or Linux with Homebrew: brew install anomalyco/tap/opencode.
Now close the terminal and open a new one. Otherwise it won’t find uv or opencode. Check that both answer:
uv --version
opencode --version
4. The practice repo
git clone https://github.com/IgnacioLD/agentic-ai-course-practice
cd agentic-ai-course-practice
uv sync
uv run pytest
uv sync installs Python and every dependency at the project’s exact versions, so we all have the same setup. The last line of uv run pytest has to say 2 failed, 8 passed.
Those two red tests are on purpose: fixing them with an agent is the first exercise. If that’s what you see, you’re all set.
This is the English version of the repo: same exercises as the Spanish taller-ia-agentica-practicas used in the room, with the code in English.
If something fails
| Symptom | What to do |
|---|---|
command not found for uv or opencode |
Open a new terminal |
uv sync hangs or gives a network error |
Try again on another connection |
| On Windows, nothing behaves like this guide | Check you’re in the Ubuntu (WSL) terminal, not PowerShell |
| Anything else | Message me on the members’ Telegram or come 30 minutes early |
See you at the first session.
Drafted with AI, edited by hand. I publish it when it sounds like me.