Set up your AI

This is the step that changes everything. Up until now, you've been setting up your workspace. Now you're adding an AI assistant — one that can read your code, write new code, explain errors, and build features from a plain English description.

A few years ago, this would've meant learning to code from scratch. Today, you describe what you want, and the AI builds it.

There are dozens of AI coding extensions, but these three stand out: they're backed by the biggest AI labs, they work as chat-based agents where you describe what you want in plain English, and they work with plans you may already have — ChatGPT Plus, Claude Pro, or Copilot Pro — so you don't have to worry about usage much.

Claude Code

Codex

Github Copilot

This landscape moves fast. These are our picks as of early 2026 — check for updates if you're reading this later.

ExtensionStyle
CodexHands-off — give it a task, come back to a finished result. Starts at $20/mo (ChatGPT Plus)
Claude CodeCollaborative — works with you interactively, step by step. Starts at $20/mo (Claude Pro)
GitHub CopilotBudget-friendly — $10/mo Pro plan gives you access to multiple models. Free tier available but limited models

You can't go wrong. Pick whichever matches a plan you already pay for, or start with Copilot Pro if you want the cheapest option.

If you want to stay absolutely free, try Antigravity — Google's VS Code fork that gives you plenty of Gemini access at no cost.

Open the Extensions panel in VS Code — click the blocks icon in the sidebar, or press ⇧⌘X · Ctrl+Shift+X . Search for the one you want, click Install, and follow the sign-in prompts.

Once signed in, open the chat panel — press ⌥⌘B · Ctrl+Alt+B or click the top-right sidebar button.

AI Chats

If VS Code asks you to trust the publisher, click Trust Publisher & Install.

Trust publisher dialog

If you see tabs like CODEX, CLAUDE CODE, or CHAT at the top of the panel, you're all set.


Try it out

In that chat panel you just opened, ask:

What is this project about?

It should summarize your project by reading through the files. If it does, your AI assistant is working and ready to help.

You've got your workspace, your local server, and your AI copilot. Time to actually use them.