IntelligentHQ
IntelligentHQ
Skip to content
Cloud application Developer tools and GitHub

GitHub Copilot

GitHub Copilot is GitHub's AI assistant for software work: code suggestions as you type, chat, agents that plan and change code and open pull requests, and AI code review. It runs in editors, the terminal, GitHub.com, GitHub Mobile and a desktop app, and is billed through plans that include GitHub AI Credits.1

Intended for
Software developers, Engineering managers and tech leads, GitHub organization and enterprise administrators, Students, teachers and open source maintainers
Runs on
SaaS on GitHub.com, IDE extensions and plugins, Command line (Copilot CLI), Desktop app for macOS, Windows and Linux, GitHub Mobile
Official Microsoft product page (opens in a new tab) Official documentation (opens in a new tab)

At a glance

What is it?
GitHub Copilot is an AI assistant that works inside GitHub and inside common developer tools. It suggests code while you type, answers questions about files and repositories, works through multi-step tasks as an agent and reviews pull requests. GitHub lists VS Code, Visual Studio, JetBrains IDEs, Xcode, Neovim and Eclipse among supported editors, plus the Copilot CLI, GitHub Mobile, GitHub.com and the GitHub Copilot app for desktop. What a user can do depends on the plan, the client and the organization's policies. Copilot is not available for GitHub Enterprise Server.123
What does it do?
Copilot cloud agent takes a task from the agents panel, a chat, an issue or a pull request, works in an ephemeral cloud environment, edits files, runs tests and linters, and opens a pull request for review. Agent mode in VS Code and the Copilot CLI do similar work locally. Copilot code review comments on pull requests. Teams can add custom instructions, prompt files, agent skills, custom agents and MCP servers. Third-party coding agents such as Claude by Anthropic and OpenAI Codex can also be used; GitHub's plans page marks third-party agents as public preview.234
Who needs it?
Individual developers who want help writing and understanding code; engineering teams that want to hand scoped tasks such as bug fixes, tests and documentation to an agent and review the result as a pull request; and organizations that need central policy control, seat management and usage budgets, which come with the Business and Enterprise plans.3
What does it need?35
  • A GitHub account
  • A Copilot plan: Free, Student, Pro, Pro+ or Max for individuals, or Business or Enterprise through an organization
  • GitHub Enterprise Cloud for Copilot Enterprise
  • Repositories hosted on GitHub for Copilot cloud agent
  • For Business and Enterprise users of the GitHub Copilot app: an administrator must enable Copilot CLI in policy settings
What it is not
Copilot does not replace review or testing: GitHub says agent session logs are no substitute for your own review. The cloud agent changes one repository and one branch per task, with a 59-minute session limit, and needs repositories hosted on GitHub. Copilot is not offered for GitHub Enterprise Server. Copilot code review for Azure Repos is a separate Azure DevOps public preview billed through Azure. Copilot Autofix belongs to GitHub Advanced Security, not to a Copilot plan.246

Availability and licensing. Plans checked 26 Oct 2026 (no effective date on the page): Free; Student (free, verified students); Pro $10/month (1,500 AI credits incl. flex); Pro+ $39/month (7,000); Max $100/month (20,000); Business $19 per seat/month (1,900 credits per user); Enterprise $39 per seat/month (3,900 per user). 1 credit = $0.01. Not available for GitHub Enterprise Server. Copilot app: preview 14 May 2026, GA 17 Jun 2026, all plans from 7 Jul 2026.3578

The problem it solves

Much developer time goes to routine work: boilerplate, tests, small fixes, documentation and reading unfamiliar code. Review queues and switching between the editor, the terminal and GitHub add further delay.

GitHub Copilot puts AI help at each of these points: suggestions in the editor, chat about the repository, an agent that turns an issue into a pull request, and automated review comments. The work stays in GitHub's normal flow of branches, checks and pull requests, so the team's existing review and merge rules still apply.

How it works

Copilot connects a model to the context and tools it needs, inside the client you use.

  • Clients: editors (VS Code, Visual Studio, JetBrains, Xcode, Neovim, Eclipse), the Copilot CLI, GitHub.com, GitHub Mobile and the GitHub Copilot app for macOS, Windows and Linux.
  • Models: each plan includes a set of models. Paid individual plans add model choice and premium models; Free and Student use automatic model selection.
  • Cloud agent: runs in an ephemeral environment that can be customized with a copilot-setup-steps.yml file, creates a branch, commits and opens one pull request per task.
  • Context and tools: repository content, custom instructions, agent skills, custom agents and MCP servers, including the GitHub MCP server.
  • Billing: usage is measured in GitHub AI Credits (1 credit = $0.01 USD). On paid plans, code completions and next edit suggestions are not billed in credits; chat, the CLI, the cloud agent, Spaces, Spark and third-party coding agents draw on them.

For Business and Enterprise, included credits are pooled across the billing entity, reset at 00:00 UTC on the first day of each month and do not carry over. Administrators can cap spend with user, cost center, organization and enterprise budgets.349

GitHub Copilot architecture: components by layer and how they connectCopilot & businessappsAgents & automationAI models & platformDeveloper toolsSecurity, identity &governanceDeveloper: Starts tasks, reviews diffs and merges pull requestsDeveloperClients: IDEs, Copilot CLI, GitHub Copilot app, GitHub.com, GitHub Mobile: Where suggestions, chat and agent sessions runClients: IDEs, Copilot CLI,GitHub Copilot app,…Copilot cloud agent: Plans and changes code in an ephemeral environment and opens a pull requestCopilot cloud agentCopilot code review: Comments on pull requests with suggested changesCopilot code reviewModels (automatic selection or chosen per plan): Generate suggestions, answers and code changesModels (automatic selectionor chosen per plan)MCP servers, skills and custom agents: Add tools and context beyond the repositoryMCP servers, skills andcustom agentsGitHub repository, branches and pull requests: Holds the code; changes pass through checks and reviewGitHub repository, branchesand pull requestsPolicies, budgets and GitHub AI Credits: Controls which features and models are allowed and caps spendPolicies, budgets and GitHubAI Credits
Diagram as a list

  1. Copilot & business apps

    • DeveloperStarts tasks, reviews diffs and merges pull requestsConnects to Clients: IDEs, Copilot CLI, GitHub Copilot app, GitHub.com, GitHub Mobile
    • Clients: IDEs, Copilot CLI, GitHub Copilot app, GitHub.com, GitHub MobileWhere suggestions, chat and agent sessions run

  2. Agents & automation

    • Copilot cloud agentPlans and changes code in an ephemeral environment and opens a pull request
    • Copilot code reviewComments on pull requests with suggested changesConnects to GitHub repository, branches and pull requests

  3. AI models & platform

    • Models (automatic selection or chosen per plan)Generate suggestions, answers and code changes

  4. Developer tools

    • MCP servers, skills and custom agentsAdd tools and context beyond the repository
    • GitHub repository, branches and pull requestsHolds the code; changes pass through checks and reviewConnects to Copilot code review

  5. Security, identity & governance

    • Policies, budgets and GitHub AI CreditsControls which features and models are allowed and caps spend
Components and connections as documented by Microsoft.

Capabilities

  • Code completion and next edit suggestions19

    Suggests code as you type; you review and accept or reject each suggestion.

    Why it matters: Speeds up routine code without leaving the editor.

    Limits: Copilot Free is limited to 2,000 completions per month. On paid plans these suggestions are not billed in AI credits.

  • Chat and agent mode19

    Answers questions about files, repositories and concepts, and in agent mode works through multi-step goals in the workspace.

    Why it matters: Gives answers grounded in your code and lets the developer delegate small multi-file changes.

    Limits: Draws on AI credits. Available features depend on the plan, the client and organization policies.

  • Copilot cloud agent4

    Researches a repository, plans, edits files, runs tests and linters in an ephemeral cloud environment and opens a pull request.

    Why it matters: Teams can hand scoped tasks to an agent and review the result through normal pull request checks.

    Limits: One repository and one branch per task, a 59-minute maximum session, repositories must be on GitHub, and branch rules can block it unless Copilot is added as a bypass actor.

  • Copilot code review12

    Reviews code and pull requests with summaries and suggested improvements.

    Why it matters: Gives authors a first pass of feedback before human reviewers start.

    Limits: Review requested by users without a Copilot license is billed in AI credits and is off by default.

  • GitHub Copilot app and Copilot CLI58

    A desktop app for macOS, Windows and Linux that runs parallel agent sessions across repositories, each on its own branch and worktree, with canvases, scheduled cloud automations and MCP tools; plus Copilot in the terminal.

    Why it matters: Manages several agent tasks outside the editor and keeps each change reviewable as a pull request.

    Limits: Business and Enterprise users need an administrator to enable Copilot CLI in policy settings. The app can also run with your own model provider key.

  • Customization and MCP124

    Custom instructions, reusable prompts, agent skills, custom agents, the MCP Registry and custom MCP servers.

    Why it matters: Aligns output with team conventions and connects Copilot to outside tools and data.

    Limits: By default the GitHub MCP server used by the cloud agent only sees the agent repository; wider access must be configured deliberately.

Practical use cases

A backlog of small bugs, missing tests and documentation tasks never gets done.
Approach
Assign well-scoped issues to Copilot cloud agent and review the pull requests it opens.
Role of GitHub Copilot
Copilot makes the first implementation pass; developers review, request changes and merge.
Data, infrastructure and skills
A paid individual plan or Copilot Business or Enterprise, repositories on GitHub, and rulesets that let Copilot push.
Type of benefit
Developer time saved
Caveats
One repository and one branch per task, a 59-minute session cap, and every change still needs human review.
First step
Pick five self-contained issues, assign them to Copilot and measure the review effort per pull request.
Pull requests wait a long time for a first review.
Approach
Turn on Copilot code review so each pull request gets automated comments before human review.
Role of GitHub Copilot
Copilot provides a first pass; human reviewers keep approval rights.
Data, infrastructure and skills
Copilot licenses for authors, or an organization decision to pay AI credits for unlicensed users.
Type of benefit
Faster feedback
Caveats
Review for unlicensed users is billed in AI credits and is off by default; it does not replace required human approvals.
First step
Enable it on one active repository and track how many comments lead to changes.
A team plans work in Azure Boards but keeps code on GitHub, and hand-offs are slow.
Approach
Connect Azure Boards to GitHub with the GitHub App and start Copilot from a work item to create a branch and draft pull request.
Role of GitHub Copilot
Copilot implements the work item; Azure Boards tracks its status.
Data, infrastructure and skills
GitHub repositories (not Azure Repos), GitHub App authentication and an active Copilot subscription.
Type of benefit
Less hand-off overhead
Caveats
A started run cannot be cancelled, and Copilot does not handle dependencies between work items.
First step
Connect one Azure DevOps project to one GitHub repository and try a small user story.

Who uses it

Works with

Optional integration

Optional integration for

Relationship labels follow Microsoft's documentation. "Alternative approaches" does not mean one is better: each profile says when it fits.

Getting started

  1. Choose a plan

    Compare Free, Student, Pro, Pro+, Max, Business and Enterprise on GitHub's plans page. Organizations that need central policy and seat control use Business or Enterprise.

    Check: Assigned users show a Copilot seat in the organization settings.

  2. Set policies and budgets

    Decide which features, models and third-party agents are allowed, and set user or organization budgets for usage beyond the pooled monthly credits.

    Check: Policy and budget settings are saved at organization or enterprise level.

  3. Install in the editor

    Install Copilot in VS Code, Visual Studio, a JetBrains IDE or another supported editor and sign in with GitHub.

    Check: Inline suggestions appear while typing in a code file.

  4. Try the cloud agent

    Assign a small issue to Copilot. Add a copilot-setup-steps.yml file if the build needs setup, and add Copilot as a ruleset bypass actor if branch rules block it.

    Check: Copilot opens a pull request linked to the issue.

Official resources

Deciding whether you need it?

Three free tools apply Microsoft's own published guidance to your situation, and say when GitHub Copilot or any new technology is not needed.

Sources

Each statement above links to the source it comes from. Labels say who reported it.

  1. About GitHub Copilot (opens in a new tab) GitHub · Vendor-reported · link checked 27 Oct 2026
  2. GitHub Copilot features and pricing page (opens in a new tab) GitHub · Vendor-reported · link checked 27 Oct 2026
  3. Plans for GitHub Copilot (opens in a new tab) GitHub · Vendor-reported
  4. About GitHub Copilot cloud agent (opens in a new tab) GitHub · Vendor-reported
  5. GitHub Copilot app available to all (changelog, 7 Jul 2026) (opens in a new tab) GitHub · Vendor-reported
  6. Copilot code reviews for Azure Repos: public preview (Azure DevOps blog, 26 Aug 2026) (opens in a new tab) Microsoft · Vendor-reported
  7. GitHub Copilot app is now available in technical preview (changelog, 14 May 2026) (opens in a new tab) GitHub · Vendor-reported
  8. GitHub Copilot app generally available (changelog, 17 Jun 2026) (opens in a new tab) GitHub · Vendor-reported
  9. Usage-based billing for organizations and enterprises (opens in a new tab) GitHub · Vendor-reported
  10. Use GitHub Copilot with Azure Boards (opens in a new tab) Microsoft · Vendor-reported

Fill out the form below to request your copy.

Name(Required)