GitHub Chat
GitHub · Coding
GitHub Chat is the AI assistant built into GitHub. It answers questions about your code, repositories, and pull requests on github.com, running on models from OpenAI, Anthropic, Google, and xAI.
About GitHub Chat
What Is GitHub Chat
GitHub Chat is the conversational side of GitHub Copilot and works as an AI coding assistant inside the browser. It lives inside the website, so you can open it from any page and ask about your own repositories. You get answers about a function you are reading, a bug you are chasing, or a library you have never touched.
The feature solves a specific problem: context switching. Instead of copying code into a separate tab and pasting results back, you ask in place and the assistant already knows which repository and file you mean. That saving matters most when you are skimming unfamiliar code or reviewing someone else's pull request.
The catch is scope. GitHub Chat only reads what your account can already see, and its answers are suggestions rather than verified fixes. You still have to check the output before you merge or ship anything. Chat usage also draws on a credit allowance, so heavy days can run through a metered plan faster than you expect.
Getting Started
- Sign in to GitHub and open github.com/copilot, or click the Copilot icon in the top bar of any page.
- Pick a model from the dropdown at the bottom of the chat panel if the default does not suit the task.
- Type a question. On a repository or pull request page, the assistant picks up that context automatically.
- Review the answer, and use the retry control to regenerate it with a different model if you want a second opinion.
Product Information
A quick look at GitHub Chat's pricing, supported platforms, and performance.
Best for
The users, tasks, and scenarios where this tool fits best.
Users
- Developers exploring an unfamiliar codebase
- Reviewers handling pull requests
- Students learning a language or framework
- Team leads triaging issues
Tasks
- Explaining a confusing file
- Generating a quick script or snippet
- Summarizing a pull request
- Answering general coding questions
- Drafting documentation from code
Scenarios
- Triage during a busy sprint
- Onboarding to a new project
- Spot-checking a dependency
- Pre-review on a large diff
Key features
Built Into Every GitHub Page
Chat opens from the top bar of github.com, and it knows which repository or pull request you are viewing. You do not paste code or explain where you are. That in-place context is the main reason to use it over a separate chatbot tab.
Automatic Repository Context
When you ask from inside a repository, the assistant pulls in the relevant files and history on its own. You spend less time describing the problem and more time reading the answer. The downside is that it can miss files outside the current view.
Model Picker
A dropdown at the bottom of the panel lets you switch Copilot chat models from OpenAI, Anthropic, Google, and xAI during a conversation. Different models cost different amounts of AI credits, so the choice affects both quality and spend. You can also retry a prompt on another model to compare answers.
Custom Instructions
You can set standing instructions that shape how the assistant responds across sessions. Teams use this to enforce tone, format, or a preferred coding style. It keeps answers consistent without repeating yourself every time.
Pull Request and Diff Summaries
On a pull request, Chat summarizes what changed and points at parts worth a closer look. It is a fast way to orient yourself before reading a long diff line by line. Treat the summary as a starting point, not a substitute for the review.
MCP Server Support
Chat connects to Model Context Protocol servers, which extend what the assistant can reach beyond the repository. That opens the door to external tools and data sources. Setup takes some configuration, so it fits teams more than casual users.
Credit-Based Usage
Chat runs on GitHub AI Credits shared across Copilot features. The free plan allows limited chat and agent use, while paid plans add a monthly base allowance. Watching your credit spend matters once you lean on the assistant daily.
Pros and cons
Pros
- Works where the code lives, so there is no copy-paste between a chat tab and your editor.
- Reads repository context automatically, which cuts the time spent explaining your problem.
- Model picker spans OpenAI, Anthropic, Google, and xAI without leaving GitHub.
- Free plan gives limited chat and agent access, enough to test the workflow.
- Custom instructions keep answers consistent across a team.
Cons
- Chat draws on metered AI credits, so heavy use on a small plan runs out quickly.
- Answers can be wrong or outdated, which means you still verify every suggestion before shipping.
- Only sees what your account can access, so private dependencies or external docs stay out of reach unless you wire them in.
- The deep features, like MCP servers and custom agents, need setup that casual users may not want to tackle.
Frequently asked questions
It answers questions about your code, repositories, and pull requests from inside GitHub. Typical uses include explaining a file, summarizing a pull request, and drafting snippets or documentation.
Related content
Explore related tools, skills, and articles for GitHub Chat.
GitHub Chat Alternatives
Base Code
Base44 (Wix) · Coding · ProductivityBase Code is a shared cloud development environment from Base44. Teammates describe changes to an existing GitHub codebase in chat, preview them live and send them back as pull requests.
Google EmbeddingGemma 2
Google DeepMind · Writing · Image · CodingGoogle EmbeddingGemma 2 is a 740M-parameter open model that turns text, code, images, audio, and video frames into vectors in one 768-dimension space, running on phones and laptops.
GitGlow
GitGlow · Coding · ProductivityGitGlow is a local-first desktop Git client that lets independent developers review their own code and their AI agents' changes in one place before shipping.