Using Copilot Chat in VS Code

The GitHub Copilot Chat extension gives you a chat interface that lets you interact with GitHub Copilot and receive answers to coding-related questions directly within VS Code, without requiring you to navigate documentation or search online forums.

Copilot Chat might use syntax highlighting, indentation, and other formatting features to add clarity to the generated response. Depending upon the type of question from the user, the result can contain links to context that Copilot used for generating a response, such as source code files or documentation, or buttons for accessing VS Code functionality.

Copilot Chat integrates in your developer flow and gives you assistance where you need it:

• Start an inline chat conversation directly from the editor for help while you're coding
• Use the Chat view to have an AI assistant on the side to help you at any time
• Launch Quick Chat to ask a quick question and get back into what you're doing

You can use GitHub Copilot Chat in various scenarios, such as:

• Answering coding questions on how to best solve a problem
• Explaining someone else's code and suggest improvements
• Proposing code fixes
• Generating unit test cases
• Generating code documentation

Getting started

Note: After reviewing this topic, you can get started with the introductory Copilot Chat tutorial.

• To use GitHub Copilot, you must have an active subscription for GitHub Copilot in your personal account, or you need to be assigned a seat by your organization.

• To use GitHub Copilot in VS Code, you must have the GitHub Copilot extension. When you install this extension, the GitHub Copilot Chat extension is also installed.

Follow these steps to set up GitHub Copilot in VS Code by signing up for a subscription and installing the Copilot extension in VS Code.

Inline chat

A key functionality of Copilot is answering questions inline, as you're coding. This allows you to harness the power of AI while staying in your existing editor workflow.

In any file, you can press ⌘I (Windows, Linux Ctrl+I) on your keyboard to bring up Copilot inline chat. You can ask Copilot questions that emerge as you write and iterate on code, such as "Explain this piece of code", or "How do I add functionality to do X?".

If you have code selected in the editor, Copilot scopes your question to the selection.

Depending on your question, Copilot can also suggest code modifications. Copilot gives a preview of the updates, which you can then accept (⌘Enter (Windows, Linux Ctrl+Enter)) or discard (Escape). Optionally, you can generate a new code suggestion.

With the inlineChat.mode setting, you can configure how inline chat shows updates in the editor. The default setting value is live, which means that the updates are applied directly in the editor, and you can compare the changes by using the inline diff viewer. If you prefer to preview the changes side-by-side, you configure the setting to preview.

The following example shows the preview mode for inline chat.

Chat view

The Chat view enables you to have a chat conversation with Copilot in a separate view. You can access the Chat view via the Activity Bar or by pressing ⌃⌘I (Windows, Linux Ctrl+Alt+I):

Copilot suggests potential questions to get you started, such as "/fix the problems in my code". You can select any of these suggestions, or use the chat input field to type your own chat prompt:

Tip: you can drag the Chat view to another location, or even open it as an editor. Learn more about custom layouts in VS Code.

Copilot Chat can provide rich and interactive results that contain: simple text, images, buttons for invoking VS Code commands, references URIs or editor locations, or file trees (for example, to show a workspace preview when a chat participant proposes to create a new workspace).

The following example shows a chat prompt to create an Express app, which returns a tree view with a suggested workspace structure and a button to create a new workspace.

As you continue asking questions, Copilot maintains the history of your conversation, and provides related follow-up questions or commands in its response too.

Tip: To get help about GitHub Copilot and how to interact with Copilot Chat, you can type /help in the chat input box.

Code blocks

Depending on your question, Copilot Chat might return source code in a code block.

Hovering over the code block presents options to Copy and Insert at Cursor (⌃Enter (Windows, Linux Ctrl+Enter)).

The More Actions (...) button gives options to Insert Into New File and Insert into Terminal (⌃⌥Enter (Windows, Linux Ctrl+Alt+Enter)).

If Copilot Chat detects that a code block contains a command, you can run it directly in the integrated terminal with Insert into Terminal (⌃⌥Enter (Windows, Linux Ctrl+Alt+Enter)). This option creates or opens the active terminal and inserts the command text, ready for you to run.

You can control the font for code blocks in chat with the following settings:

• chat.editor.fontFamily
• chat.editor.fontSize
• chat.editor.fontWeight
• chat.editor.lineHeight

Tip: Navigate between code blocks with Chat: Next Code Block (⌥⌘PageDown (Windows, Linux Ctrl+Alt+PageDown)) and Chat Previous Code Block (⌥⌘PageUp (Windows, Linux Ctrl+Alt+PageUp)).

Quick chat

If you want to ask Copilot a quick question and don't want to start a full Chat view session or open inline chat in your editor, you can use the Quick Chat dropdown. To open Quick Chat, you can run Chat: Open Quick Chat in the Command Palette, or use the ⇧⌘I (Windows Ctrl+Shift+I, Linux Ctrl+Shift+Alt+I) keyboard shortcut.

You can type questions, scope your questions with chat participants and slash commands, and promote the discussion to a full Chat view session with the Open in Chat View button in the upper right of the dropdown.

Multi-turn conversations

When you ask a question in Copilot Chat, you aren't stuck with the first response. Keep the chat conversation open and continue to iterate and prompt Copilot to improve the suggested solution. Copilot has both the context of the generated code and your current conversation history. As you keep asking additional questions, Copilot further refines the response according to your requirements.

Here's an example using inline chat to create a function to calculate Fibonacci numbers:

If you prefer a solution that doesn't use recursion, update the chat prompt and submit it to Copilot to get a different response.

Learn more about how to iterate over a Copilot chat conversation.

Chat smart actions

Smart actions are integrated in your VS Code flow (for example, in Quick Fix and context menus) and they do not require you to write any prompt at all. To make it easier to use Copilot Chat features, there is a Copilot menu group in the editor context menu. Right-click in the editor and navigate to Copilot to see the available options.

The most powerful smart action is /fix. Here is a relatively simple TypeScript calculator with an error noting that "Argument of type 'string' is not assignable to parameter of type 'number'". Select the sparkle icon, and then choose Fix using Copilot.

Similar to /fix, the /doc smart action is popular with users. To use /doc, select a block of code, right-click, and choose Copilot > Generate Docs. Copilot will generate a documentation comment for your code.

Chat participants

To further help Copilot give you more relevant answers, you can indicate the scope and intent of your question through chat participants.

Chat participants are like experts who have a specialty that they can help you with, and you can talk to them in the chat by mentioning them with the @ symbol. Currently, there are the following built-in chat participants:

• @workspace has context about the code in your workspace and can help you navigate it, finding relevant files or classes.
• @vscode knows about commands and features in the VS Code editor itself, and can help you use them.
• @terminal has context about the integrated terminal shell and its contents.

You can prepend your chat inputs with a specific participant to help Copilot give you a more relevant response.

@workspace

The @workspace chat participant knows how to gather context about the code in your workspace, can help you navigate it, find relevant classes, files, and more. Imagine you are in the VS Code repository and you want to find out more about the service in charge of the current ICodeEditor; you can use the chat participant like this:

Because it has all of the necessary context, @workspace can answer the kinds of questions that developers are much more likely to ask. For example, questions that pertain to how different parts of the code interact:

• "@workspace how are notifications scheduled?"

Or questions that require knowledge of related code, dependencies, and design patterns:

• "@workspace add form validation, similar to the newsletter page"

@vscode

VS Code can be customized in so many ways that people get pleasantly surprised when they discover some hidden functionality. To help you to unlock the full power of VS Code, we created @vscode.

This chat participant knows all about VS Code and can help you bridge the gap between natural language and VS Code commands and customizations. @vscode internally uses tools that give it access to the index of all the settings and commands and we are in the process of adding a tool so that this chat participant can also use the VS Code documentation. Now you can ask vague questions like:

• "@vscode the name of that thing when vscode fake opens a file? And how to disable it?"
• "@vscode how do I change my VS Code colors?"

@terminal

With @terminal you can ask questions about the integrated terminal shell, its buffer, and the current selection. In the following example, you use the Quick fix Explain using Copilot in the terminal to get information about a failed shell command.

Notice how the prompt in the Chat view is populated with @terminal #terminalLastCommand to help correct the error. #terminalLastCommand is a chat variable, which allows you to add more context to the chat prompt. Learn more about chat variables.

Note: the terminal doesn't pull in the workspace context, which could take some time.

Slash commands

Chat participants can also contribute what we call slash commands, which are shortcuts to specific functionality provided by the chat participant. One of the tasks when answering questions is to determine the intent, understanding what you want to do.

For example, we can infer that "Create a new workspace with Node.js Express Pug TypeScript" means that you want a new project, but "@workspace /new Node.js Express Pug TypeScript" is explicit, concise, and saves you time typing.

Once the intent is clear, @workspace has a better chance of addressing your needs, despite the inherent ambiguity of natural language. The @workspace chat participant can propose a directory structure and users can select the proposed files to preview them. There is a Create Workspace button that generates these files in a new folder.

Examples of built-in commands:

• /help: get help about using GitHub Copilot
• /doc: generate code documentation
• /clear: start a new chat session
• @workspace /explain (or /explain): explain how the selected code works
• @workspace /fix (or /fix): propose a fix for the problems in the selected code
• @workspace /tests (or /tests): generate unit tests for the selected code
• @workspace /new (or /new): scaffold code for a new workspace
• @workspace /newNotebook (or /newNotebook): create a new Jupyter Notebook
• @vscode /api (or /api): ask about VS Code extension development
• @vscode /search (or /search): generate query parameters for the Search view
• @terminal /explain: explain terminal functionality or shell commands

Improve the performance of Copilot Chat

Copilot works best when it has sufficient context to know what you're doing and what you want help with. Just as you would provide a colleague with the context when asking for help with a specific programming task, you can do the same with Copilot.

You can help Copilot provide better answers by upvoting or downvoting responses with the thumbs up and down icons in the upper right of its response. This provides Copilot feedback on how much it helped with your scenario so that it can help you even better in the future.

Use chat variables

Chat participants, such as @workspace or @vscode, can contribute chat variables that provide domain-specific context. You can reference a chat variable in your chat prompt by using the # symbol. For example, the #selection variable contains the text selection in the active editor.

By using a chat variable, you can be more specific about the context that you include in your chat prompt. For example, the prompt "which sorting algo is used #selection" focuses the chat request on the selected code snippet.

Examples of built-in chat variables are:

• #selection: the visible source code in the active editor
• #editor: the current selection in the active editor. The editor content is implicitly included in the Chat view context.
• #file: include a specified file in your workspace as context with your chat prompt.
• #terminalSelection: the active terminal's selection
• #terminalLastCommand: the active terminal's last run command

Note: if possible, the full contents of the file will be included when you use #file. If that is too large to fit into the context window, an outline of the file will be included that includes functions and their descriptions without implementations. If the outline is also too large, then the file won't be part of the prompt.

Use chat participants and commands

Chat participants, such as @workspace, might have context associated within them. For example, @vscode is aware of VS Code settings and its APIs. When you include a chat participant in the chat prompt, you inherently add context to the prompt. Similarly, by using slash commands, you can further focus the intent of the chat request.

Chat history

Copilot keeps track of the history of your conversation in the Chat view. Copilot can use this information as context in subsequent prompts, for example when you have a multi-turn conversation.

For example, when you first ask "what is the number data type in TypeScript?" and then ask "Can you use this type for decimal numbers?", Copilot knows that you're still referring to the number data type. Notice that the suggested prompt is also relevant within our conversation.

You can delete a prompt and the corresponding response from the chat history by hovering over the prompt and selecting the x control. It might be useful to delete one or more prompts to get more relevant responses.

You can export all prompts and responses for a chat session in a JSON file with the Chat: Export Session... command (workbench.action.chat.export) in the Command Palette.

Context in inline chat

If you're using inline chat, and you have code selected in the editor, Copilot scopes your question to the selection.

Ignore files

When you're using @workspace to ask questions related to your VS Code workspace, you can explicitly ignore specific files from the workspace by using a .gitignore file. @workspace respects .gitignore when deciding which files from the workspace to index.

Use voice interactions

With the voice control capabilities in VS Code, provided by the VS Code Speech extension, you can initiate a chat conversation by using your voice:

• Use your voice to dictate your chat prompt
• Use the "Hey Code" voice command to start a voice session with Copilot Chat
• Accelerate voice input for chat by using the "hold to speak" mode

Learn more about how to use voice interactions in VS Code.

Privacy and transparency

We emphasize responsible usage of AI, especially when it comes to source code. The inlineChat.acceptedOrDiscardBeforeSave setting, which is enabled by default, asks you for confirmation before saving code that was generated by Copilot.

When the setting is enabled, a file save operation waits for you to accept or discard any pending inline chat session. This also applies when Auto Save is enabled, which will be temporarily disabled until inline chat has ended.

To enable more workspace search features for private repositories, we require additional permissions. If we detect that we don't have these permissions already, we will ask for them at startup. Once granted, we'll securely store the session for the future.

Learn more about security, privacy, and transparency in the GitHub Copilot Trust Center.

Additional resources

You can read more about GitHub Copilot and how to use it in VS Code in the GitHub Copilot documentation.

Or check out the VS Code Copilot Series on YouTube, where you can find more introductory content and programming-specific videos for using Copilot with Python, C#, Java, PowerShell, and more.

Next step

Get started with the introductory Copilot Chat tutorial.