What is GitHub Copilot?
GitHub Copilot is an AI coding assistant that works inside your code editor, offering inline completions, a chat sidebar, and an agent mode that can make multi-file changes. Built by GitHub (owned by Microsoft), it runs in VS Code, Visual Studio, JetBrains IDEs, Neovim, and on github.com.
It suggests code as you type, answers questions about the code you're looking at, generates tests, and in agent mode plans and applies changes across several files. On paid plans you can pick the model behind chat and agents, choosing among frontier models from Anthropic, OpenAI, and Google. A separate coding agent can take an assigned GitHub issue and open a pull request you review.
If you want AI help where you already write code, completing a line, explaining a file, or handing off a task, Copilot is the most established option.
What it's best for
- In-editor completion: context-aware suggestions as you type, accepted with a keystroke.
- Staying in your IDE: VS Code, Visual Studio, JetBrains, and Neovim, with no switch to a separate chat tab.
- Asking about your codebase: chat that sees your open files and repo to explain, fix, and write tests.
- Agentic, multi-file changes: agent mode edits across files, and the coding agent works asynchronously on issues.
- Choosing the model for the job: paid plans let you switch between frontier models for chat and agents.
- Pull-request workflows on GitHub: summaries and review suggestions in the place your code already lives.
Where it falls short
- Non-developers. It's a coding tool, not a general-purpose assistant.
- Open-weight or fully offline use. It's a hosted service tied to your GitHub account.
- Teams that forbid any code leaving their environment without a vetted enterprise configuration.
- Buyers who want a single bundled price rather than per-seat or per-request billing on heavier tiers.
Getting started
Install the GitHub Copilot extension in your editor (or use it directly on github.com) and sign in with your GitHub account. The free tier offers a limited number of completions and chat messages per month; paid plans remove those limits.
Once active, suggestions appear inline as you type. Press Tab to accept, or keep typing to ignore them.
Completions and chat
Inline completion is best when you give it a head start: a clear function signature, a descriptive name, or a comment describing intent. Copilot Chat answers questions about selected code and the wider repo.
Slash commands speed up common asks (/explain a confusing file, /tests to generate unit tests, /fix to address an error), all grounded in the code you have open.
Agent mode and the coding agent
Agent mode takes a higher-level prompt and makes multi-step edits across several files, showing you a summary to review before you accept.
The Copilot coding agent goes further: assign it a GitHub issue and it works asynchronously, then opens a pull request for your team to review, useful for well-scoped, self-contained tasks.
Getting better answers
Keep the relevant files open so Copilot has the context it needs, and write the comment or signature before asking for a completion.
Review every suggestion before accepting, because you own the code that ships. Treat Copilot as a fast junior pair, not an authority, especially on security-sensitive code.
What GitHub Copilot costs
Approximate, in USD, as of January 2026. Prices change often. Confirm on the official site before you rely on them.
Free
$0
A limited number of completions and chat messages per month for individuals.
Pro
~$10 / month
Unlimited completions and chat, model choice, and agent mode.
Pro+
~$39 / month
Highest limits and the most generous access to the most capable models.
Business / Enterprise
~$19 to 39 / user / month
Org policy controls, IP indemnity, audit, and (Enterprise) knowledge grounded in your repos.
Example prompts
Copy these into GitHub Copilot as starting points, then adapt them to your task.
Generate tests
/tests: write unit tests for the selected function covering the happy path, error cases, and edge cases. Match the project's existing test framework and style.
Explain unfamiliar code
/explain this file as if I'm new to the codebase: what it does, the key types, and where the side effects and external calls are.
Hand off a scoped task
Add input validation to every endpoint in this router, return a 400 with a typed error on failure, and update the tests. Show me a summary of the files you changed.
GitHub Copilot
common questions.
Direct answers to the questions we get asked the most. If yours isn't covered, write to the team.