Coding

GitHub Copilot: completion inside the editor

How Copilot's inline suggestions and chat work, where it helps most, and how to review what it writes.

  • Admin
  • 3 min read

Pricing

Free tier with limits; paid plans from about US$10/month (check the official page)

Official site

Open the official website (opens in a new tab)

Prices and plans change. Check the official page before you pay.

Copilot puts a model in your editor. The difference from a chat window is context: it sees the file you are in, the function above, and often the tests beside it, which makes its suggestions more specific than a prompt typed into a browser.

What it is good at

  • Boilerplate. A getter, a loop, a test skeleton, a switch over an enum.
  • Continuing a pattern. Write two similar cases and it proposes the third in the same style.
  • Tests. Given a function, it writes plausible cases, including boundaries you might not think of.
  • Explaining unfamiliar code. Select a block, ask what it does, get a plain-English answer in place.

Where it fails

It completes what looks right, which is not the same as what is correct. The common failures:

  • A plausible API that does not exist, or has a different signature in your version.
  • Code that compiles and quietly mishandles an edge case.
  • Suggestions that duplicate a helper your project already has.
  • Tests that assert the implementation rather than the behaviour, so they pass while the bug remains.

The review discipline is the same as for a human pull request: read what it wrote, run it, and check the boundary cases.

Inline completion versus chat

Inline suggestions are for the next few lines; accept them with Tab and keep moving. Chat is for a question about the current file, a refactor across several files, or an explanation.

The productive rhythm is to write the signature and the test first, then let the model fill the body. You keep the design decisions and delegate the typing.

Prompting in the editor

  • Name the file and the intent: "add pagination to this endpoint, matching the style of listUsers".
  • Reference existing code: "use the same error handling as parseOrder".
  • Ask for a diff, not a rewrite, when the change is small.
  • Keep the selection tight. A whole file as context produces a whole-file rewrite.

Pricing

There is a free tier with a monthly cap on completions and chats, and paid individual plans from around US$10 per month. Business plans add policy controls and per-seat billing. Student and open-source programmes exist; check the current terms.

The licence question

Suggestions can resemble public code. GitHub’s filters reduce verbatim matches, and organisations can enable settings that block suggestions matching public repositories. For a commercial project, know your organisation’s policy before relying on a filter.

Who it is for

Developers who already know the language. Copilot amplifies fluency: it saves typing for someone who can spot a wrong suggestion, and it teaches bad habits to someone who cannot yet tell.

Alternatives

Cursor and Windsurf are editors built around a model, with multi-file edits as the primary workflow. Claude Code works from the terminal. The model matters less than the workflow: choose the one that fits how you already work.

Admin

About the author

Admin

Instructor and technical writer at tutorials.in

Writes beginner lessons for tutorials.in. Each lesson is checked against the language or tool documentation, and examples are run before publishing.

Reviewed by the editorial team before publishing. Last updated November 16, 2026.