copilot-chat

Subscribe to all “copilot-chat” posts via RSS or follow GitHub Changelog on Twitter to stay updated on everything we ship.

~ cd github-changelog
~/github-changelog|main git log main
showing all changes successfully

A dark-themed code editor interface displaying a preview of a 'Planetary Travel Time' calculator. The interface includes a rocket icon, a title, and a description prompting users to select a planet and a NASA spacecraft to calculate travel time from Earth. Dropdown menus show 'Mars' as the selected planet and 'Voyager 1' as the selected spacecraft. A floating 3D GitHub Copilot assistant with a glowing, futuristic helmet is visible in the bottom-right corner.

Inspired by our previous release, working with Copilot Chat on GitHub has become even more seamless. You can instantly preview HTML files, edit files you’ve created, and work on issues right away. Several exciting new capabilities give you more control and flexibility.

What’s new

  • Preview your rendered HTML files directly in the side panel
  • Edit files in the side panel to seamlessly refine and adjust them
  • Generate and preview Mermaid diagrams for fast visualizations, whether they’re flowcharts or sequence diagrams
  • Keep tabs on your issues in the same right side panel, ensuring you can tackle open tasks while discussing them
  • Track issues or pull requests in responses that are rendered in a familiar GitHub style, making working with them easier

In addition, you can enjoy a smoother streaming experience and enhanced rendering of attachments.

Try it out

See the updated experience in action by submitting any of the following example prompts:

Join us as we continue to streamline Copilot Chat, giving you instant previews, flexible editing, and more power right where you need it! Your feedback drives our improvements. Let us know how these new changes enhance your workflow by using the in-product feedback option or sharing your thoughts in the GitHub Community.

See more

GitHub Copilot now features instant semantic code search indexing, dramatically reducing the time it takes for Copilot to understand and reference your codebase.

What’s changed

Previously, when you wanted GitHub Copilot to reference your repository’s code in its responses, the semantic code search indexing process would take approximately five minutes to complete. With this update, indexing now completes in just a few seconds in most cases, though it may take up to 60 seconds. This means you can get contextually-aware Copilot assistance almost immediately after opening a repository.

Why this matters

Semantic Code Search is one of GitHub Copilot’s most powerful capabilities, enabling GitHub Copilot to:

  • Provide responses specific to your codebase’s architecture and patterns.
  • Reference existing functions, classes, and implementations in your repo.
  • Suggest code that aligns with your project’s style and conventions.
  • Answer questions about your codebase with accurate, context-aware information.

With instant semantic code search indexing, there’s virtually no waiting period between opening a repository and receiving codebase-aware AI assistance, making your development workflow more efficient and interruption-free.

How it works

Semantic code search indexing is automatically triggered when you open GitHub Copilot Chat on github.com. For VS Code users with the GitHub Copilot extension, you can also manually trigger indexing through the Copilot UI if needed.

Availability

This feature is available to all GitHub Copilot users across all tiers, including the free tier. There are no limits on how many repositories can be indexed.

Learn more

For detailed information about repository indexing for GitHub Copilot, check out our documentation.

Join the discussion within GitHub Community.

See more

Personal instructions in github.com

Personal instructions are now generally available for Copilot Chat on github.com! This means you can provide Copilot with important details about your preferences, such as your preferred language, response style, or even code standards.

To get started, open up Copilot Chat, click ..., and select Personal instructions. That’s it! Copilot will now incorporate your preferences for all chats in github.com.

💡 Looking for ideas? Here are some examples to kick things off:

  • Frameworks: “Use Vue 3 with the composition API.”
  • Language: “Always respond in Portuguese.”
  • Response preferences: “Each message should communicate a singular idea. That idea should be limited to ONE codeblock or paragraph.”
  • Code style and best practices: “Optimize code for readability and performance. Utilize SOLID principles.”
See more

You can now attach images and work with them directly in Copilot Chat in VS Code or Visual Studio. Share screenshots of errors and Copilot will interpret the image and resolve the issue. Or share mockups of new designs, and Vision will help you bring them to life.

You can now attach images using several methods:

  • Drag and drop images from your OS or from the Explorer view
  • Paste an image from the clipboard
  • Attach a screenshot of the editor window (in VS Code, select Attach > Screenshot Window)

Currently, the supported image types are JPEG/JPG, PNG, GIF, and WEBP.
You can use the vision capability with the GPT-4o model. Make sure you have the Copilot ‘Editor Preview Features’ policy enabled to get access.

This feature was previously available to VS Code Insiders and Visual Studio Preview users, and is now available to users of the stable editor versions in public preview.

Copilot Chat analyzing an image

To learn more, read the documentation about using Vision in Copilot Chat.

Please share your feedback in our community discussions.

See more

GitHub Copilot GPT-4.5

OpenAI’s latest model, GPT-4.5, is now available in GitHub Copilot Chat to Copilot Enterprise users. GPT-4.5 is a large language model designed with advanced capabilities in intuition, writing style, and broad knowledge. It performs effectively with creative prompts and provides reliable responses to obscure knowledge queries. GPT-4.5 will launch in Visual Studio Code and on github.com for Copilot Enterprise users with a limit of 10 requests every 12 hours per user. In the coming weeks, we’ll be scaling rate limits and extending support to Visual Studio and JetBrains.

GPT 4.5 in the VS Code Model Picker

As model releases have continued to accelerate, we’ve been thinking about how we can sustainably offer advanced AI models like GPT-4.5 to more GitHub users. This includes individual developers who want the most advanced capabilities from day one. Stay tuned for updates.

Enabling access

Copilot Enterprise administrators will need to enable access to GPT-4.5 via a new policy in Copilot settings. As an administrator, you can confirm availability by checking your individual Copilot settings and confirming the policy for GPT-4.5 is set to “enabled”. Once enabled, users will see GPT-4.5 in the Copilot Chat model selector in VS Code and on github.com.

See more

A stylized image showing a 3D Copilot head looking at a user message that has been edited via the new editing functionality

The immersive mode of Copilot chat on GitHub now provides you with increased control and flexibility at every step.

What’s new:

  • ✏️ Edit your previously sent messages: Refine your messages whenever you need
  • 🔄 Reload responses: Get different variations instantly
  • 📝 Keep track of all iterations in one place: Easily view the history of your message edits and responses
  • 🔍 Compare different message and response pairs seamlessly: Switch between versions to see which one works best

Pro tips:

  • 🛠️ Fix typos or add context without starting over
  • 🧪 Experiment with different phrasings to get the best responses
  • 🔀 Toggle between versions to compare outcomes

These new refining capabilities make it easier to iterate with Copilot. Make adjustments and improvements while maintaining the full context of your conversations, helping you to stay focused when chatting with Copilot chat.

We hope these changes help keep you in the flow state when chatting with Copilot on GitHub.com 💫.

💬 Let us know what you think using the in-product feedback option or pop it into the GitHub Community at any time.

See more

Copilot Extensions GA

Your tools. Your workflows. All within Copilot Chat.

GitHub Copilot Extensions are now generally available for users across all Copilot license tiers. With Copilot Extensions, you can integrate and prompt your favorite tools directly in Copilot Chat using natural language wherever you develop, including Visual Studio Code, Visual Studio, JetBrains IDEs, and GitHub.com. Copilot Extensions on GitHub Mobile will be generally available in the coming weeks.

Copilot Extensions help you stay in your workflow, with context-aware assistance from your favorite tools right at your fingertips. Today’s marketplace is home to a wide range of extensions, from Perplexity to Stack Overflow, to Docker and Mermaid Chart. Developers can unlock productivity gains with extensions in minutes. For example, Arm’s extension streamlines cloud adoption and migration, enabling developers to build, test, and deploy software on Arm-based servers while seamlessly leveraging Arm’s efficient, scalable, and high-performance architecture.

Explore these extensions and more on the GitHub Marketplace to bring new contexts and capabilities into the chat. All you need is access to GitHub Copilot to get started. 🚀

Building GitHub Copilot Extensions

Our platform also empowers you to build your own public or private extension depending on your requirements. This flexibility allows you to develop extremely customized extensions for your enterprise or organization, or develop general applications that can serve thousands of developers. The comprehensive Copilot Extensions toolkit provides you with centralized code samples and tools to help you build high quality extensions.

Alongside General Availability, we’re introducing OpenID Connect (OIDC) support for builders. This replaces the X-Github-Token auth model with native third-party tokens, reducing API round trips, and improving security. Instead of verifying GitHub tokens on every request, integrators receive pre-exchanged tokens tailored to their system, enabling direct authentication and authorization. This lowers latency, simplifies identity mapping, and aligns with GitHub’s existing OIDC workflows for Actions.

Builders have several ways to develop customized extensions, including:

  • Copilot skillsets, a faster, lightweight implementation option
  • Context passing, a capability that helps extensions benefit from a user’s local editor context for more tailored responses

Ready to contribute to our growing ecosystem? Get started with our Copilot Extension builder docs.

👀 What’s next?

Our general availability is only the starting point for agentic capabilities. We’re continuing to reimagine AI assisted workflows, with recent releases like agent mode and explorations around Project Padawan. These innovations only scratch the surface of what is possible with GitHub and AI agents. Continue being a part of the conversation by providing feedback as you try out extensions. ⭐

See more

Personal custom instructions are now available in public preview for Copilot on github.com 🎉

Copilot Chat on github.com now supports personal instructions! This means you can provide Copilot with important details about your preferences, such as your preferred language, response style, or even code standards.

To get started, open up Copilot Chat, click ..., and select Personal instructions. That’s it! Copilot will now incorporate your preferences for all chats in github.com.

💡 Looking for ideas? Here are some examples to kick things off:

  • Language preferences: “Always respond in Portuguese.”
  • Response preferences: “Be concise and to-the-point. Always cite your sources.”
  • Personal preferences: “You are a seasoned React developer with ten years of experience.”
  • Code preferences: “Always provide examples in TypeScript.”

Looking for more? Visit prompts.chat for more inspiration.

Search the web 🔍 in Copilot chat using Bing

Copilot Chat can now search Bing to answer questions and find information beyond its general knowledge or your codebase. This feature makes it easy to chat about recent events, trends, and new developments. It’s now generally available in VS Code, Visual Studio, and github.com.

Since our public preview, we’ve enhanced our web search capabilities to deliver more relevant and accurate responses to your questions.

Give it a try today:

  • “What’s the latest version of React?”
  • “What are the most recent updates in Python’s machine learning libraries?”

Quality improvements to Copilot Chat

In the last month, we have delivered a collection of quick wins and subtle enhancements ✨, making the product smoother and more delightful for everyone.

What’s new:

  • Better search in Copilot Chat for more relevant and complete answers, including expanded lexical search results and larger semantic search responses for better context
  • Improved memory in Copilot, now keeping more of your chat history for better context and flow
  • Improved Copilot Chat’s awareness of the README.md when asking about a repository
  • Enhanced Copilot’s awareness of time, showing times relative to you instead of UTC

We also recently added the ability to view and generate new code files to Copilot chat at github.com/copilot – check it out!

See more

OpenAI’s latest model, o3-mini, is now available in GitHub Copilot Free.

o3-mini model picker GIF

OpenAI’s latest model, o3-mini, is the most cost-efficient model in their reasoning series. o3-mini outperforms o1 on coding benchmarks with response times that are comparable to o1-mini, meaning you’ll get improved quality at nearly the same latency. The model is configured to use OpenAI’s medium reasoning effort and can be accessed in VS Code and Github.com Copilot chat today, with support to follow soon in Visual Studio and JetBrains.

Access to o3-mini is currently in preview and is subject to the 50 free chats per month limit when using Copilot Free.

Get started with Copilot Free on GitHub and in VSCode today or learn more in our documentation.

See more

Announced at GitHub Universe 2024, Gemini 2.0 Flash is now available to all GitHub Copilot customers, including Copilot Free. The latest Gemini model from Google shows high capabilities for code suggestions, documentation, and explaining code.

Get started today!

Copilot Free or Pro users

You can start using the new Gemini 2.0 Flash model today via the model selector in Copilot Chat in Visual Studio Code and immersive chat on GitHub.com.

Copilot Business or Enterprise users

Copilot Business and Enterprise organization administrators will need to grant access to Gemini 2.0 Flash in Copilot via a new policy in Copilot settings. Once enabled, you will see the model selector in VS Code and chat on GitHub.com. You can confirm availability by checking individual Copilot settings and confirming the policy for Gemini 2.0 Flash is set to enabled.

Share your feedback

Join the community discussion to share feedback and tips.

For additional information, check out the docs on Gemini 2.0 Flash in Copilot.

See more

A screenshot showing the new file blocks in the Copilot chat feed and the new side panel where code files are opened.

You can now view and iterate on generated files directly within Copilot chat on GitHub. File suggestions open in a side panel, making it faster than ever to preview and refine your generated code.

What’s new:

  • Quick file iteration: When Copilot chat suggests files, they open immediately for closer inspection.
  • Convenient previews: Markdown files render directly inside the side panel.
  • Support for more formats: We’re just getting started, expect additional file types and capabilities soon.

Try it out:

Your feedback drives our improvements! Let us know what you think using the in-product feedback option or share your thoughts in the GitHub Community.

Join us on this journey as we continue to enhance Copilot Chat and deliver a smoother developer workflow!

See more

Copilot and GitHub Models o3-mini Release

OpenAI’s latest model, o3-mini, is now available in GitHub Copilot and GitHub Models, bringing OpenAI’s newest reasoning model to your coding workflow.

The o3-mini reasoning model outperforms o1 on coding benchmarks with response times that are comparable to o1-mini, meaning you’ll get improved quality at nearly the same latency.

This cutting-edge model is rolling out gradually and will be available to GitHub Copilot Pro, Business, and Enterprise users today via the model picker in Visual Studio Code and github.com chat (support in Visual Studio, and JetBrains are coming soon). To accelerate your workflow, whether you’re debugging, refactoring, modernizing, testing, or just getting started, simply select “o3-mini (Preview)” to begin using it.

Paid Copilot subscribers get up to 50 messages every 12 hours. Business or Enterprise admins can enable o3-mini access for org members through their org and enterprise admin settings pages.

GitHub Models users with a paid Copilot plan will also be able to leverage the o3-mini model to enhance their AI applications and projects later today. In the GitHub Models playground, you can explore o3-mini’s versatility as you experiment with sample prompts, refine your ideas, and iterate as you build. You can also try it alongside other models available on GitHub Models including models from Cohere, DeepSeek, Meta, and Mistral.

To learn more, check out product documentation on GitHub Models. You can also join our community discussions.

See more

Repository administrators and organization owners with a Copilot Business or Copilot Enterprise license can use content exclusions to configure Copilot in GitHub.com to ignore certain files. For example, ignore files called “secrets.json”, ignore files with the “*.cfg” extension, or ignore all files in the “/scripts/**” directory. Content exclusions can be applied through Settings > Copilot > Content exclusion.

See more

Copilot chat on GitHub.com just leveled up! It now supports repository-specific custom instructions, giving you a more tailored and personalized chat experience. This means you can provide Copilot with important details about your development stack, coding standards, or even how you prefer to chat. 🎉

Now, you can use repository-specific custom instructions when chatting with Copilot in Visual Studio, VS Code and on GitHub.com

🚀 Getting Started

  1. Create a .github/copilot-instructions.md file in your repository (if .github doesn’t exist yet, go ahead and create it).
  2. Add your custom instructions to the file.

That’s it! Copilot chat will automatically apply these instructions whenever you’re chatting about that repository.

💡 Looking for ideas? Here are some examples to kick things off:

  • JavaScript: “Omit semicolons in code examples.”
  • Python: “We use Poetry for dependencies, not pip. Share instructions using Poetry.”
  • General style preference: “Use arrow functions instead of traditional function expressions.”

Start customizing and make Copilot chat feel like an extension of your team! 🛠

See more

On March 31, 2025, GitHub Copilot Extensions will require an updated header format for agent requests. Both updated and previous versions of the request headers will be supported until then. These headers denote requests that come from GitHub and enable your extension to communicate with GitHub.

Updated headers:
X-GitHub-Public-Key-Identifier
X-GitHub-Public-Key-Signature

Previous headers, to be deprecated on March 31, 2025:
Github-Public-Key-Identifier
Github-Public-Key-Signature

Please update your relevant checks to the correct headers by March 31, 2025 for a consistent experience and to avoid breaking changes. To learn more, visit this page.

See more