GitHub Copilot

GitHub Copilot is GitHub’s AI developer platform, integrated across the software development lifecycle. It began as a pair-programmer for code completion and has grown into a broader suite built around a multi-model approach: developers can choose the underlying LLM to fit the task — fast reasoning, large context, or creative generation.

Multi-model choice

Copilot Chat lets you switch models rather than lock into one — for example Anthropic’s Claude 3.5/3.7 Sonnet for complex coding, Google’s Gemini 1.5 Pro for very large context windows (up to 2M tokens), or OpenAI’s o3-mini for high-speed reasoning.

Copilot Workspace (agentic workflow)

A task-centric environment that takes a GitHub Issue to a Pull Request in one flow: it analyzes the issue, proposes a natural-language plan, generates a spec, and implements the code across multiple files. If the plan is off, edit the natural-language steps and the AI regenerates the implementation to match.

Context and completion

• Next Edit Suggestions (NES): predicts where you are likely to edit next and suggests the change, speeding up refactoring.
• Repository indexing: for Enterprise, Copilot indexes the whole organization’s codebase and answers architecture questions (“How does our authentication service talk to billing?”) with knowledge of proprietary and legacy code.

Chat and extensions

• Extensions platform: Copilot Chat supports third-party extensions, so you can query external tools from the IDE (“Check Sentry for recent errors in this function,” “Deploy this to Azure”).
• Inline chat: highlight code and ask Copilot to refactor, debug, or explain it without leaving the editor.

Copilot Autofix

Integrated with GitHub Advanced Security, it detects vulnerabilities such as SQL injection or XSS and generates a pull request to fix them.

Supported environments

• IDEs: VS Code, Visual Studio (2022/2026), JetBrains (IntelliJ, PyCharm), and Xcode with full Chat support.
• CLI: the gh copilot extension generates shell commands and explains terminal output.
• Mobile: Copilot Chat and agent sessions via the GitHub mobile app.

Where it fits

Legacy modernization through Workspace, incident response via monitoring-tool extensions, test-driven development (write tests first, generate the implementation), and cross-language porting that respects the target language’s idioms.

Direct link: https://github.com/features/copilot