How to Learn GitHub Copilot

To learn GitHub Copilot, follow the quickstart for GitHub.com. Copilot is built into GitHub, so you can use it without installing anything. The quickstart takes about ten minutes: open a file, ask what it does, then assign an issue to Copilot cloud agent and review the pull request it opens. Where to use GitHub Copilot says the same: start on the website.

About GitHub Copilot groups the work into four kinds: assistive suggestions and answers you review one by one, agentic tasks you still approve, customizations such as instructions, skills, and custom agents, and coding agents, models, and tools from other providers, including MCP servers.

Plans for GitHub Copilot include Copilot Free: limited access to a selection of features, at no cost. The GitHub.com quickstart says Copilot cloud agent in that tutorial needs a paid plan, and it names Copilot Pro. Without a paid plan, that page sends you to the IDE quickstart for the remaining steps.

Ask about a file

Open a file in any repository. Click the Copilot icon at the top right of the file view, type a question, and press Enter. The quickstart's examples are:

• Explain what this file does.
• How could I improve this code?
• What tests would cover this function?
• What was the last change and why was it made?

Copilot answers in the chat panel. Ask Tell me more to expand the last comment. You can also ask exploratory questions about the repository, ask what changed in a pull request, or ask about a specific issue or commit. More examples are in Asking GitHub Copilot questions in GitHub.

Assign an issue to Copilot cloud agent

Assigning an issue is how the quickstart starts Copilot cloud agent, and Copilot opens a pull request when it finishes.

In a personal repository you can write to, enable Copilot cloud agent from your Copilot settings. Open an issue, or create one. In the right sidebar, click Assignees, then Copilot. An optional prompt can name coding patterns, files to modify, or testing requirements. Copilot uses the issue title, the description, and the comments already there. Add later notes on the pull request after it opens.

Read the diff and check that it matches the issue. To request a revision, mention @copilot in a pull request comment, or push commits to the branch. Merge when the change is the one you wanted. The quickstart points to Review output from Copilot.

Suggestions, chat, and agent mode in the IDE

The IDE quickstart is the writing path. It covers Visual Studio Code, Visual Studio, JetBrains IDEs, Eclipse, and XCode. You need an active Copilot plan, the extension for that IDE, and a GitHub sign-in in the editor.

Open a file, then open Copilot Chat. In Visual Studio Code the shortcut is Ctrl+Alt+I on Windows or Linux, and Control+Command+I on macOS. Elsewhere, Copilot Chat is in the menu bar or sidebar. Type what does this file do. Select a line and type explain this line. Copilot can see the open file, so the question can point at that code.

Most IDEs support agent mode. Confirm yours on the Copilot feature matrix. In chat, select Agent and describe the task. The example is Create a task manager web app with the ability to add, delete, and mark tasks as completed. Copilot chooses the files, makes the edits, and runs commands with your approval. Review the changes, accept them in the IDE, and commit when they look right.

For an inline suggestion, create a JavaScript file and type:

function calculateDaysBetweenDates(begin, end) {

Copilot suggests a function body in grayed text. The suggestion may vary. Press Tab to accept it. The plans page limits inline completions on Copilot Free to 2000 a month.

Best practices use inline suggestions for a snippet, repeated code, or a test, and chat for a question or a larger block you will revise. Read a suggestion before you keep it. Open the files the task needs and close the rest. Accept or reject an inline suggestion, and use thumbs up or thumbs down on a chat reply.

The CLI and the Copilot app

Every plan on the plans page includes Copilot CLI and the Copilot app. Getting started with GitHub Copilot CLI installs with npm install -g @github/copilot (Node.js 22 or later), winget install GitHub.Copilot on Windows, or brew install --cask copilot-cli on macOS or Linux. In the project, run copilot, then /login, and confirm that you trust the files there. Copilot will not change those files without your explicit approval. Ask Give me an overview of this project. Outside a session, pass the prompt with copilot -p, and add -s for the response alone.

The GitHub Copilot app is the desktop place for agent sessions, parallel tasks, and the pull request lifecycle. Sign in, connect a local folder or a GitHub repository, and start a session in a mode such as Interactive. The quickstart's prompt asks for a small low-risk change, the implementation, and an explanation of the diff. Open Changes, then Create PR if you want to keep it. The app does not support GitHub Enterprise Server.

A course from the quickstart

Practice is matching the surface to the job: a question on an open file, a cloud-agent pull request you still review, Tab for an inline suggestion, Agent in the IDE, then the CLI or the app.

Ailurn turns a prompt, a PDF, a GitHub repo, or a docs URL into a course you take in the same workspace. It does not watch YouTube videos, and pasting a link does not import a transcript. It does not issue accredited certificates. File attachments start on a paid plan. Start from the AI course builder and ask for a course on the GitHub.com Copilot quickstart: explain a file in chat, assign an issue to Copilot cloud agent on a paid plan, review the pull request with @copilot, then practice inline suggestions and agent mode in the IDE.