GitHub Copilot Chat is now generally available for Visual Studio Code and Visual Studio, and ships with every GitHub Copilot plan alongside the original code completion capabilities. Verified teachers, students, and maintainers of popular open source projects can use it at no cost.
Access for individuals and organizations
Copilot Chat is available to all GitHub Copilot individual users. Enterprise and organization administrators grant access to their development teams by enabling the Copilot Chat setting for their users. Teams already using Copilot Chat in beta, or already granted access, need to take no further action.
GPT-4 and natural-language prompting
The assistant runs on GPT-4 and is built as a contextually-aware AI assistant for developer scenarios. It accepts prompts in the natural language of the user's choice—German, Portuguese, and others included—and returns real-time guidance inside the IDE. Capabilities span explaining complex coding concepts, detecting security vulnerabilities, and writing unit tests.
Because Copilot Chat is personalized to each developer's coding practices, where and when to use it is a matter of individual workflow. GitHub engineers have shared their own usage patterns in the following areas.
Translating code between programming languages
There are many reasons why I need to translate between programming languages, including standardization efforts, breaking up monoliths, and looking for performance benefits. Copilot Chat makes it easier than ever to do exactly that by helping me understand the code that needs to be translated, and then actually translating that code into the target language. This not only saves me lots of time rewriting code and transferring logic, but also reduces cognitive load and mental effort so I can focus on quality and design.
I recently had to do this for a demo where I translated a JavaScript Application to Java, and then again to C#.
Given some fundamental differences between the languages and the frameworks, the first step was bootstrapping an initial version of the app in the target language and framework. With Copilot Chat I used the @workspace /new-Slash-Command to have this up and running within a minute. After that, a simple prompt like `Translate this File to java/C#` would often be enough for Copilot Chat to present a high quality result, automatically picking up the traits of the target language and framework. Finally, to ensure quality and performance, I used the @workspace /test-slash command to generate unit tests with ease.
Answering coding questions in real time
I routinely find myself opening up Copilot Chat to quickly turn an idea I have into a bash command that gets me the information I’m looking for, and saves me from having to stop what I’m doing and look up specific command flags that I routinely forget. Recently, I opened Copilot Chat to convert a helper function I had been using in my Ruby tests to valid Typescript so I could use it in part of a new project I’m working on.
Being able to quickly pull up Copilot Chat, ask a question, and tweak the results to get exactly what I’m looking for has been a welcome change from hunting through pages of search results.
Refreshing coding practices on demand
I've worked on many different tech stacks and products over my career, and have held several different roles with many teams at GitHub, too. What happens to me more than I'd care to admit now is looking at the screen and waiting for the fingers to start typing, because I know I've done this before but I'm struggling to remember how exactly. Where I used to try to hack at it again, search my history, or just search in general, now I can bring up Copilot Chat to rapidly jog my memory, and maybe even learn a few new tricks on top of that. I still enjoy pairing with others and collaborating on the big things because that's where strong relationships develop and big growth happens, but using Copilot Chat to reach what's just out of reach in memory or get me started on the basics of something new is so helpful.
Looking at my most recent interchanges, I've blasted through translating complex SQL syntax to KQL, generating a wild regexp, testing an API scenario in my Codespace, and … oh, wow, I asked it to explain Dependabot as a pirate so I could answer my teenage son's question and make him laugh at the same time. I appreciate that it's always there and keeps me going, so I can focus more on the big thinking.
GitHub positions Copilot Chat within its broader effort to bring AI to the entire development lifecycle. A getting-started guide for using Copilot Chat in your IDE is available in the GitHub documentation.



