Setting Up Windows for Agentic AI Programming

 in AI  tagged Windows / setup / installation / AI / agentic by

This post summarizes and adapts setup notes from Ed Donner’s Udemy course The Complete Agentic AI Engineering Course (2025)

These instructions — refined collaboratively by Ed Donner, ChatGPT, and Mike Porter — are designed to help learners get their local development environment running smoothly. If you’re setting up for the first time, take your time and follow along carefully — these steps will prepare you to explore, experiment, and build Agentic systems with confidence.

Setting Up Your Development Environment (for Windows PC)

Working at the cutting edge of AI requires a reliable and well-configured environment. This guide will walk you through the steps needed to get up and running smoothly with Ed Donner’s course. While setup on Windows can have a few tricky moments, we’ve included key tips to make your journey as painless as possible.

Before You Begin — Important Notes

Many students run into trouble with a few common Windows limitations.
Please take the time to check these before moving forward — doing so will save you headaches later!

  • 1. Permissions — Review this tutorial to ensure you have proper Windows permissions.
  • 2. Anti-Virus, Firewall, and VPN — These can sometimes block installations or network access. Temporarily disable them if necessary.
  • 3. The 260-Character File Path Limit — Windows can choke on long file paths. Fix this by following these instructions and restarting your computer.
  • 4. Microsoft Build Tools — Required for many Data Science packages. Install them using these steps, or refer to this guide if you’re on Windows 11.

Part 1 — Clone the Repository

You’ll begin by downloading the course repository to your local machine.

  • Install Git for Windows using the default options.
  • Open Command Prompt (Win + R → type cmd → Enter).
  • Navigate to your project folder, for example:
    cd C:\Users\YourUsername\projects
    or
    cd D:\MyData\AI\projects or wherever you store your projects
  • If you don’t have a projects folder, create one:
    mkdir C:\Users\YourUsername\projects
  • Clone the course repo:
    git clone https://github.com/ed-donner/agents.git
  • Move into the new folder:
    cd agents

Part 2 — Install Cursor (or VS Code)

Cursor is a powerful AI-enhanced IDE built on VS Code. You can use either Cursor or standard VS Code for this course.

  • Visit cursor.com and create an account.
  • Download and install Cursor.
  • Open a new window and choose “Open Project.” Navigate to the agents directory you cloned earlier.
  • Install the recommended extensions for Python and Jupyter if prompted. Otherwise, manually install them via View → Extensions.
  • Open the Explorer pane (View → Explorer) to browse course files by week.

Part 3 — Install uv (Fast Python Package Manager)

uv is a modern, high-performance package manager that’s replacing pip and conda in many Data Science setups. It’s fast, reliable, and easy to use.

  • Install uv using the standalone installer at Astral Docs.
  • In Cursor, open a Terminal (View → Terminal) and confirm you’re in the agents directory (pwd).
  • Run uv self update to ensure you’re on the latest version.
  • If you’ve used Anaconda, deactivate it first:
    conda deactivate
    and disable auto-activation:
    conda config --set auto_activate_base false
  • Then run:
    uv sync — it will automatically install Python 3.12 and all required dependencies.
  • If you see SSL errors, try:
    uv --native-tls sync or uv --allow-insecure-host github.com sync
  • Install CrewAI tools:
    uv tool install crewai
    uv tool upgrade crewai

Confirm Installation

  • You should now have a .venv folder inside agents.
  • uv python list should show Python 3.12.x.
  • uv tool list should show CrewAI.

Quick tip: with uv you can replace pip install package with uv add package, and python script.py with uv run script.py. The tool updates and syncs automatically, so you rarely need to call uv sync manually.

Part 4 — Add Your OpenAI API Key (Optional)

If you want to use OpenAI’s GPT models directly (recommended for full performance), you’ll need an API key.
If you prefer free alternatives like Ollama or OpenRouter, you can skip this section.

  • Create an account at OpenAI Platform.
  • Add a small credit (e.g., $5) and disable auto-recharge in Billing Settings.
  • Generate a new API key at OpenAI API Keys and save it securely (starting with sk-proj-).
  • Later you may also add:

Part 5 — Create the .env File

This file securely stores your API keys so they can be read by your Agentic apps.

  • In Cursor, create a new text file and add your keys. Example:
OPENAI_API_KEY=sk-proj-yourkeyhere
GOOGLE_API_KEY=xxxx
ANTHROPIC_API_KEY=xxxx
DEEPSEEK_API_KEY=xxxx
  • Save the file in the agents directory and name it exactly .env (not env.txt or .env.txt).
  • Double-check that the key names are typed correctly (e.g., OPENAI_API_KEY, not OPEN_API_KEY).

You’re Ready to Go!

In Cursor, open the 1_foundations folder and double-click 1_lab1.ipynb.
When prompted to select a kernel, choose .venv (Python 3.12.x).
If you don’t see it, adjust your VS Code Settings → “Path to folder with a list of Virtual Environments” to point to your project folder (e.g., C:\Users\username\projects\agents).

If you hit a snag, open troubleshooting.ipynb for help or reach out to Ed at ed@edwarddonner.com.

Final Thoughts

That’s it — you’ve just built a professional-grade AI environment! With Git, Cursor (or VS Code), uv, and CrewAI in place, you’re ready to start exploring the world of Agentic AI systems.

Created collaboratively by Ed Donner, ChatGPT, and Mike Porter.

Leave a Reply