GitHub Copilot
AI coding assistant from GitHub that provides completions, chat, agent mode, code review, and a CLI across major IDEs and GitHub.com.
At a Glance
A fast way to get started with GitHub Copilot for individual developers, with no credit card required.
Engagement
Available On
Alternatives
Updated Nov 2026
About GitHub Copilot
GitHub Copilot is GitHub's AI coding assistant. It runs in the editor, in the terminal, in a desktop app, and across GitHub.com and GitHub Mobile. It suggests code, answers questions, edits multiple files, runs agent tasks, and reviews pull requests. It offers models from Anthropic, Google, Microsoft, Moonshot AI, OpenAI, and xAI. It is sold as individual and organization plans, and usage is metered in GitHub AI Credits.
Where It Runs
- Editor integration — Available in VS Code, Visual Studio, JetBrains IDEs, Neovim, Xcode, Eclipse, Zed, Raycast, and SQL Server Management Studio; feature support varies by editor
- Terminal — Copilot CLI is an agent in the terminal that answers questions, writes and debugs code, and creates pull requests, with interactive, plan, and programmatic modes; included on all plans
- GitHub Copilot app — A desktop app for macOS, Windows, and Linux, built on Copilot CLI, for running several agent sessions in parallel and managing issues and pull requests; available on all plans
- GitHub.com and beyond — Chat, agent tasks, pull request summaries, and code review on GitHub.com and GitHub Mobile, with Copilot also available in Slack and Microsoft Teams
Core Coding Features
- Inline suggestions — Code completions and next edit suggestions that are unlimited on paid plans, do not use AI credits, and are capped at 2,000 per month on Copilot Free
- Copilot Chat — Conversational help in IDEs, on GitHub.com, in GitHub Mobile, and in Windows Terminal, for explaining, refactoring, and debugging code
- Agent mode — Multi-step task execution in VS Code, Visual Studio, JetBrains, Eclipse, and Xcode that plans, edits across files, and runs commands
- Copilot in the cloud — Successor to the Copilot cloud agent experience, live on paid plans since October 28, 2026: assign work to Copilot to research, plan, and write code in a cloud sandbox, and pick up the same session from the web, mobile, the app, VS Code, or the CLI
- Third-party coding agents — Anthropic Claude and OpenAI Codex agents can take issues and prompts on GitHub in public preview, powered by an existing Copilot subscription
Review and Collaboration
- Code review — AI review comments and suggested fixes on pull requests, on GitHub.com, in IDEs, and on mobile, at Lite, Balanced, or Max effort levels; Business and Enterprise admins can extend it to members without a Copilot license
- Pull request summaries — Auto-generated descriptions of changes in GitHub.com pull requests
- Copilot Spaces — Shared collections of repositories, issues, pull requests, notes, images, and files that ground Copilot's answers; can be shared with a team or publicly
Memory and Context
- Copilot Memory — Generally available on paid plans since October 28, 2026; stores repository facts and personal preferences and reuses them across the Copilot app, CLI, VS Code, cloud agent, code review, Mobile, Slack, and Teams
- Copilot Dreaming — Opt-in public preview that reviews past sessions and review feedback in the background, updates memories automatically, and opens pull requests that propose changes to instruction and skill files such as AGENTS.md and .github/copilot-instructions.md; billed in AI credits; the preview requires GPT-6.1 Sol access and is not available on Copilot Pro
- Cross-repository code search — Copilot can search code, configuration, and docs across an organization's repositories without a local clone, including semantic search by behavior
Models and Extensibility
- Model selection — Model picker includes GPT-6.1 Sol, GPT-5.3-Codex, Claude Opus 5.5, Claude Sonnet 5.5, Claude Haiku 5.5, Gemini 4 Argon, Grok 4.7, Kimi K3, and MAI-Code-1.1-Flash; Free and Student plans use auto model selection only
- MCP, connectors, and plugins — Connects to Model Context Protocol servers, Copilot connectors, and plugins; admins can restrict MCP use to an allowlist or a private registry
- Custom instructions, agents, and skills — Repository, personal, and organization instructions, plus custom agents, agent skills, prompt files, and hooks to fit Copilot to a codebase
- Copilot SDK — SDKs for TypeScript, Python, Go, .NET, Java, and Rust that expose the Copilot CLI agent runtime; the SDK is being renamed Microsoft Copilot SDK, and existing users do not need to change anything
Enterprise and Trust
- Enterprise controls — User management, SAML SSO, audit logs, usage metrics, content exclusion, IP indemnity, model and agent policies, and spending controls on Business and Enterprise
- Privacy posture — GitHub may train models on interaction data only from individual Free, Pro, and Pro+ subscribers, who can opt out in settings; Business and Enterprise data is handled under GitHub's Data Protection Agreement
- Code referencing — Optional filter that blocks suggestions matching public code on GitHub, or shows reference details for matches when allowed
GitHub Copilot is built into GitHub's developer platform. GitHub replaced premium requests with usage-based AI credits for most plans in 2026. Recent work has centered on agents: Copilot in the cloud, the GitHub Copilot app, Copilot CLI, third-party agents, and MCP governance. GitHub has also added Copilot Memory and Dreaming, which carry context from one session to the next.
Community Discussions
GitHub previews Dynamic Workflows for repeatable agent orchestration
GitHub introduced Dynamic Workflows in public preview across Copilot CLI, the Copilot app and SDK. Developers write orchestration as code, with sequential or parallel agents, structured handoffs, verification steps, human checkpoints and pause/resume. Announcement. The useful distinction is who defi…
That review belongs to the old commit
An agent asks Copilot for review through the new API, changes the code, then someone reads the old review as if it covers the new patch. That’s the annoying case I’d want the workflow to prevent. Tie the review to the commit, otherwise we’ve just automated some confusion.
Which parts of your agent workflow would you put in code?
Copilot’s Dynamic Workflows let you write the steps in code and leave the reasoning to agents. That seems useful for things like “run the tests before asking for review.” Where would you draw the line? I’d want the program to decide when a change can move forward, but leave room for the agent to fig…
Pricing
Free
A fast way to get started with GitHub Copilot for individual developers, with no credit card required.
- 50 agent mode or chat requests per month
- 2,000 completions per month
- Access to Haiku 4.5, GPT-5 mini, and more
- Copilot CLI
Pro
Accelerate workflows with GitHub Copilot for individual developers who need unlimited completions and chat with leading models.
- Everything in Free
- Copilot cloud agent
- Copilot code review
- Claude and Codex on GitHub and VS Code
- 300 premium requests, with the option to buy more
- Unlimited agent mode and chats with GPT-5 mini
- Unlimited inline suggestions
- Access to models from Anthropic, Google, OpenAI, and more
Pro+
Scale with agents and more models for individual developers who want access to the full model roster.
- Everything in Pro
- Access to all models, including Claude Opus 4.7 and more
- 5x as many premium requests as Pro to use the latest models, with the option to buy more
- Access to GitHub Spark
Business
Accelerate workflows with GitHub Copilot for organizations that need policy controls, audit logs, and IP indemnity.
- Copilot cloud agent
- Copilot code review
- Claude and Codex on GitHub and VS Code
- 300 premium requests to use latest models per user, with the option to buy more
- Unlimited agent mode and chats with GPT-5 mini
- Unlimited inline suggestions
- Access to models from Anthropic, Google, OpenAI, and more
- User management and usage metrics
- IP indemnity and data privacy
- Copilot CLI
Enterprise
Scale with agents and more models for enterprises that need the full model roster and GitHub Spark access.
- Everything in Business
- Access to all models, including Claude Opus 4.6 and more
- 3.33x as many premium requests than Business to use the latest models, with the option to buy more
- Access to GitHub Spark
Capabilities
Key Features
- Inline code completions
- Copilot Chat in IDE and GitHub.com
- Agent mode with multi-file edits
- Copilot cloud agent for asynchronous tasks
- Copilot CLI for the terminal
- AI code review on pull requests
- Pull request summaries
- Copilot Spaces for shared team context
- Model selection (Anthropic, Google, OpenAI, and more)
- Copilot Extensions and MCP server support
- Custom instructions and agent skills
- Code referencing and duplication filter
- IP indemnity (paid plans)
- Audit logs and policy management (Business and Enterprise)
- Content exclusion rules
- GitHub.com integration (Enterprise)
- Access to GitHub Spark (Pro+ and Enterprise)


