Skip to content

GitHub Copilot

GitHub Copilot can suggest code as you type in a supported editor or prepare changes for review on GitHub. Choose based on how much project access the task needs and how you want to inspect the result.

Read official information

When should you use Copilot in your editor or on GitHub?

Use editor suggestions when you want to see and accept individual edits as you work. A supported editor, its Copilot integration and an eligible GitHub sign-in are the starting requirements.

A cloud task suits a change you can describe and review later. Copilot can work in a separate hosted environment and prepare a branch and pull request, a proposal to merge changes. GitHub lists cloud access for the paid individual plans Copilot Pro, Pro+ and Max. Check that your repository is eligible and that any organization or enterprise policy allows cloud work.

What can Copilot see and change?

An inline suggestion can send nearby code and context from open editor tabs to the model. Check which account is active and which files you are allowed to share before opening a project with Copilot enabled.

For a cloud task, choose only a repository you have permission to share. Feedback on an existing pull request can create commits on its branch. GitHub documents an approval step before workflows run on proposed changes by default. Inspect the changed files before authorizing a workflow or merging.

Can you check one suggested edit?

Use a fresh local folder containing only the sample below, with no remote, credentials or connections to a real app. Close other project tabs and confirm the active GitHub account is approved for this sample. If you can’t verify those conditions, compare the two expressions on paper.

function hasRoom(reservedSeats) {
  return reservedSeats < 4
}
  1. Record the starting point
    Save the sample as room.js and record a starting version with your Git client. The sample room has space for four reservations.
  2. Change one condition
    The sample room now has space for six reservations. Edit only the return line, using an inline suggestion if one appears. Check that the result is reservedSeats < 6 before keeping it.
  3. Check both sides of the limit
    Compare the changed file with the starting version. The return line should be the only change. With five reservations there should be room. With six there should be none. Check both values on paper before committing the edit.

What can GitHub Copilot charge for?

On usage-based plans, metered features such as chat and cloud agent tasks consume GitHub AI Credits according to the model and the text it processes. GitHub lists one AI credit as US$0.01 as of November 11, 2026. Ordinary editor completions and next edit suggestions are separate from AI-credit charges. The Free plan limits completions.

Paid personal plans have a monthly subscription and included usage. GitHub lists Business and Enterprise prices per granted seat per month, with organization-level credit and spending settings. If you use cloud automations, GitHub Actions minutes can be a separate cost for the repository owner.

When you start a session on GitHub in the default Cloud environment, GitHub also meters the sandbox. Its running time in seconds and the memory assigned for each second draw from the owner’s monthly compute allowance. The owner can be billed once that allowance is used, and snapshots kept after stopping can have storage charges. Check the cloud sandbox budget and whether it stops usage at its limit. The AI-credit budget is separate.

Some existing annual Pro and Pro+ subscribers stayed on older premium-request billing. Check your account’s allowance and budget before starting metered work.