← All postsHow Do You Use GitHub Copilot in IntelliJ?

How Do You Use GitHub Copilot in IntelliJ?

Carlos GarciaCarlos GarciaPublished

How Do You Use GitHub Copilot in IntelliJ?

GitHub Copilot runs inside IntelliJ IDEA through a plugin, and the install takes about two minutes.

The sign-in is the part that trips people up, because it uses a device code rather than a normal OAuth redirect.

There is a reason to read GitHub's documentation for this rather than a tutorial: the installation procedure GitHub publishes is written against IntelliJ IDEA specifically, and it says so. "The following procedure will guide you through installation of the GitHub Copilot plugin in IntelliJ IDEA. Steps to install the plugin in another supported IDE may differ."

This guide follows that procedure, then covers the question nobody answers honestly — which Copilot features actually work in IntelliJ, and which ones GitHub flags as preview or excludes outright.

What Is the Short Answer?

Open your Plugins settings, search the Marketplace for the GitHub Copilot plugin, install it, restart the IDE, then sign in through Tools and GitHub Copilot and Login to GitHub.

You need a Copilot entitlement first. GitHub states it plainly for JetBrains: "To use GitHub Copilot in JetBrains, you need either limited access through Copilot Free or a paid Copilot plan for full access."

Copilot Free is enough to try it. Note that this is one of the relatively few places where GitHub makes a JetBrains-specific statement rather than a general all-IDEs one.

Your developer tooling is sorted. Is your organic traffic? Get a free SEO audit and see what is holding your site back.

How Do You Install the Plugin?

GitHub compresses the whole navigation into a single sentence, which is worth quoting because almost every third-party guide invents a breadcrumb that GitHub never prints: "In your JetBrains IDE, open the Plugins settings, search the Marketplace for the GitHub Copilot plugin, then click Install."

So the three labels are Plugins, Marketplace and Install, in that order. GitHub does not say which menu the Plugins settings sit under on that page.

Then restart. The label is a button rather than a menu action: "After GitHub Copilot is installed, click Restart IDE."

Which JetBrains IDEs are supported

GitHub publishes fifteen: IntelliJ IDEA (Ultimate, Community, Educational), Android Studio, CLion, Code With Me Guest, DataGrip, DataSpell, GoLand, JetBrains Client, MPS, PhpStorm, PyCharm (Professional, Community, Educational), Rider, RubyMine, RustRover and WebStorm.

Worth highlighting because it is widely got wrong: IntelliJ IDEA Community and Educational editions are both on that list. Copilot in IntelliJ is not an Ultimate-only feature.

Free audit

Where do you stand in AI search?

A manual SEO + AI search audit of your site: how ChatGPT, Gemini, Perplexity and Claude read and cite it today, and what to fix first. Report within 48 hours, no credit card.

Get my free audit

Or have it all done: AI SEO services, $999 once

How Do You Sign In?

This is the device-code flow, and the labels matter because the dialogs are terse.

  1. After the IDE restarts, "click the Tools menu. Click GitHub Copilot, then click Login to GitHub."
  2. In the "Sign in to GitHub" dialog box, click Copy and Open. That single button copies the device code and opens the activation window together.
  3. A device activation window opens in your browser. Sign in with your GitHub account if prompted, or use Continue with Google or Continue with Apple.
  4. Paste the device code, then click Continue.
  5. Approve the permissions by clicking Authorize GitHub Copilot Plugin.
  6. Back in the IDE, click OK to finish.

Two labels are easy to misremember. The authorise button says Authorize GitHub Copilot Plugin — with "Plugin" on the end. And the menu entry is Login to GitHub, written as one word, which is GitHub's own copy.

If you are on a managed user account through GitHub Enterprise, there are account settings to change before this flow will work. That is a separate path and your administrator controls it.

Where Copilot's own settings live

Once signed in, the plugin's options are under the normal JetBrains settings route — and here GitHub does give a menu path: "In your JetBrains IDE, click the File menu (Windows), or the name of the application in the menu bar (macOS), then click Settings." Then "In the left sidebar click Tools, click GitHub Copilot".

Shipping faster is only half the job — people still have to find you. Request a free audit and find the gaps in your site's search performance.

Which Features Actually Work in IntelliJ?

GitHub maintains a feature matrix for Copilot across editors, and it is the most useful page here because its own legend separates two things most comparisons blur.

A tick means supported, a cross means not supported, P means under preview, and a dash means the feature does not apply to that IDE. A cross is an affirmative statement of non-support, not silence.

Against the JetBrains plugin at version 1.5.66, compared with VS Code:

Documented as fully supported in JetBrains, same as VS Code: agent mode, chat, code completion, code referencing, Copilot code review, edit mode for multi-file changes, MCP, checkpoints and workspace indexing.

That list deserves emphasis because the folklore is wrong. Agent mode is not VS Code-only — it is the default chat mode in JetBrains, which GitHub describes as the "Full agentic experience with autonomous task execution." Copilot code review and multi-file edits are documented as supported too.

MCP support reached full support at plugin version 1.5.53, having been preview before that.

Flagged as preview in JetBrains while generally available in VS Code: custom instructions, next edit suggestions, agent skills, custom agents and prompt files.

Explicitly not supported in JetBrains: the Java Upgrade Agent, which is in preview for VS Code. A genuine cross, not a silence.

Custom instructions are shaped differently here

This is the sharpest real parity gap. "JetBrains IDEs support a single .github/copilot-instructions.md custom instructions file stored in the repository, and a locally stored global-copilot-instructions.md file."

Compare that with VS Code, which GitHub says "supports three types of repository custom instructions", and Visual Studio, which "supports two types". JetBrains gets one repository file plus one global file.

The global file is JetBrains-only, and its directory is literally named intellij on both platforms: /Users/YOUR-USERNAME/.config/github-copilot/intellij/ on macOS, and C:\Users\YOUR-USERNAME\AppData\Local\github-copilot\intellij\ on Windows.

There is a contradiction inside GitHub's own documentation here, and it is worth knowing before you build a workflow on it.

The prose page documents JetBrains custom instructions as a working feature with a full procedure and no caveat, while the feature matrix marks custom instructions as preview for JetBrains. Those are two different claims about the same thing.

When Should You Use the Plugin, and When Not?

The plugin is the right choice when you want the full Copilot surface inside the IDE you already live in — completions as you type, chat in a side panel, agent mode for multi-file work.

But GitHub documents three entry points for Copilot in JetBrains, not one, and most guides mention only the first: "In JetBrains IDEs, there are three primary entry points for accessing Copilot to choose from: the GitHub Copilot plugin, the JetBrains AI Assistant, or GitHub Copilot CLI in the integrated terminal."

And crucially: "The plugin provides the full Copilot experience, but you can also access Copilot through JetBrains AI Assistant or Copilot CLI without installing the plugin."

So if you already pay for JetBrains AI Assistant, you can reach Copilot through it without a second plugin.

And if your work is mostly terminal-shaped, Copilot CLI in the integrated terminal may suit you better than an editor plugin.

Skip the plugin if you are evaluating Copilot's newest capabilities specifically. The preview-versus-GA gaps above are real, and VS Code gets those features first.

What Are the Limitations Worth Knowing?

GitHub publishes no minimum JetBrains IDE version. This is the thing to take away, because it is the single most commonly fabricated fact about Copilot in IntelliJ. There is a dedicated "Version compatibility" heading in GitHub's documentation, and its entire body is a pointer elsewhere: "For information about version compatibility of the GitHub Copilot extension in JetBrains IDEs, see GitHub Copilot Versions in the JetBrains Marketplace."

That deferral is sharper than it looks, because GitHub is plainly willing to publish a version floor when it has one. In the same prerequisites list, on the same page, Visual Studio gets "Visual Studio 2022 version 17.8 or later".

JetBrains gets a list of product names and "Latest version of the GitHub Copilot extension" — no number anywhere. If a tutorial tells you that you need IntelliJ 2021.2 or later, that figure did not come from GitHub.

The Copilot Free allowance is only half published. Completions have a hard number: "Limited to 2000 completions per month on Copilot Free." Chat does not. GitHub has moved to an AI credits model and now says only that "Copilot Free and Copilot Student both have an allowance of AI credits" — and Copilot Free is not a row in the table that lists actual credit figures. The familiar "50 chat requests per month" is from retired documentation. On free and student plans, "access to models is available through Copilot auto model selection only", and Copilot Chat skills in IDEs are listed as not available on Copilot Free.

GitHub cannot decide whether it is a plugin or an extension. On one page the heading "About the GitHub Copilot extension in JetBrains IDEs" is followed by "Installing the GitHub Copilot plugin in your JetBrains IDE". One prerequisites bullet manages both words in a single sentence. JetBrains' own UI says Plugins, and GitHub's URL anchor uses "plugin", so plugin is the safer term — but expect to search for both when you are reading the docs.

The licensing is GitHub's, not JetBrains'. Easy to miss and occasionally matters to a procurement team: "GitHub, Inc. is the licensor of the JetBrains plugin... JetBrains has no responsibility or liability in connection with the plugin or such agreement."

Copilot in IntelliJ vs the Alternatives

Against Copilot in VS Code. VS Code is where features land first, and five capabilities are GA there while still preview in JetBrains. If you are already productive in IntelliJ, that gap is not worth switching editors over — the core loop of completions, chat, agent mode and multi-file edits is documented as supported in both.

Against JetBrains AI Assistant. AI Assistant is built into the IDE and is more deeply wired into JetBrains' own refactoring and inspection machinery. Copilot's plugin gives you GitHub's model selection, repository instructions and the GitHub-side features like code review. You can run Copilot through AI Assistant, so this is less either-or than it looks.

Against Copilot CLI. If you mostly drive work from a terminal, the CLI in IntelliJ's integrated terminal avoids a plugin entirely. It is the lightest option and GitHub documents it as a first-class entry point.

Against a terminal-first agent. Tools that live outside the editor take a different approach — you give an instruction and they work across files autonomously rather than completing your next keystroke. Worth comparing if agentic work, not autocomplete, is what you are actually buying.

One more thing worth auditing this week. Get a free SEO audit and find out what your site is missing.

Final Thoughts

Installing Copilot in IntelliJ is four steps and a device code. The useful knowledge is everything around it.

Three things are worth remembering. GitHub publishes no minimum IntelliJ version, and any specific number you read elsewhere is invented — check the Marketplace versions page instead. Agent mode, code review, multi-file edits and MCP are all documented as supported in JetBrains, so do not accept the claim that they are VS Code exclusives.

And custom instructions work differently here: one repository file and one global file, against three types in VS Code, with GitHub's own matrix calling the feature preview while its prose page documents it as working.

If you are weighing Copilot against the other assistants before committing, our guide to using GitHub Copilot covers the plans, the modes and where the tool falls short across every editor it supports.

Ready to see what your site could rank for? Claim your free SEO audit and get a prioritised list of fixes.

Want this done for you?

SEO Stuff gets businesses cited by ChatGPT, Gemini, Perplexity and Claude, and ranking in Google. Start with the free audit, a call, or the package.

See where you stand

A free, manual SEO + AI search audit of your site, with a report within 48 hours. No credit card, no sales call.

Get my free audit

Talk it through

A free 15-minute call with the founder about whether AI search is a meaningful opportunity for your business.

Free strategy call

Have it done

The Done-For-You Package: audit, 10 pages of content, 3 DR50+ placements, dashboard. $999 once, delivered in 21 business days.

See the package