☕️ Welcome to the GitHub Copilot Café #175367
Pinned
HarvirChima
started this conversation in
Discover
Replies: 2 comments
-
Love this! Thanks for sharing! 🎉 |
Beta Was this translation helpful? Give feedback.
0 replies
-
Thanks @HarvirChima |
Beta Was this translation helpful? Give feedback.
0 replies
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Uh oh!
There was an error while loading. Please reload this page.
-
☕ Welcome to the GitHub Copilot Café
Step into the Copilot Café. The espresso is strong, the atmosphere is buzzing, and your AI coding companion is already a few steps ahead of your order.
At GitHub, we’re not just brewing a better editor experience. We’re shaping a new kind of developer workflow. Whether you’re stopping in for a quick espresso shot of code help or settling in for a long afternoon refactor, Copilot has something for every kind of developer.
Pull up a chair. Let’s walk through the menu.
☕ Ask Mode: Just a quick question
Ask Mode is like walking up to the counter and saying, “Can I get something quick, smooth, and no fuss?” It’s designed for those small moments when you just need clarity. Want an explanation of a regex that looks like it came from another dimension? Need a quick summary of a test suite? Curious how a particular function works?
With Ask Mode, you type the question in plain language and Copilot responds instantly in the chat panel, grounded in the context of your project. It’s the coding equivalent of a barista sliding an espresso across the counter with a smile. No overthinking. No context switching. Just fast, useful answers that keep you in flow.
You can chat with Copilot in GitHub, and if you want to understand how Ask, Edit, and Agent modes fit together, check out this GitHub Blog Post.
🥐 Edit Mode: Your usual, with a twist
Sometimes you want your regular latte, but with oat milk or an extra shot. Edit Mode is built for that. Highlight a block of code, tell Copilot what you want, and watch it transform. Whether it’s “make this loop more efficient,” “convert to arrow functions,” or “simplify this for readability,” Edit Mode makes the change while respecting your intent.
It’s like having a barista who knows your usual order but can remix it on the fly. The magic here is precision. You stay in control, Copilot makes the edit, and you can approve or refine until it’s exactly right. For quick refactors, stylistic changes, or cleanup before a commit, Edit Mode is your go-to.
Explore Copilot edits.
🧠 Agent Mode: The barista behind the counter
Now we’re moving beyond the quick orders. Agent Mode is when the barista knows your name, remembers your order, and gets started before you even say a word. Agent Mode plans, executes, and verifies tasks across your repo. It can edit multiple files, run tests, call tools like linters or analyzers, and check in with you before applying any changes.
The beauty of Agent Mode is the workflow. It doesn’t just make edits, it follows a Plan → Act → Verify loop. That means fewer mistakes, less context switching, and more time for you to focus on design and architecture. It’s like asking, “Could you reorganize the pastry display, standardize the prices, and add a seasonal drink?” and the barista rolls up their sleeves and makes it happen.
If you want to see it in action, start with the Agent Mode documentation and dive into the Agent Mode 101.
🤖 Copilot Coding Agent: A barista that runs the whole café
If Agent Mode is a skilled barista, the Copilot Coding Agent is the café manager who can run the whole shop. It doesn’t just handle a single drink order or a few refactors. It can manage larger, multi-step tasks across your project. You can say, “Refactor the user module, add validation, run the tests, and propose a PR,” and it will orchestrate the process end to end.
Built on top of Agent Mode, the Coding Agent takes autonomy a step further. It coordinates more complex workflows across your repo and tools, all while keeping you in the loop for approvals. It’s the difference between ordering a cappuccino and asking the café to cater your offsite and they say, “No problem, it’ll be ready on time.”
Learn more about the Coding Agent and the vision behind it in this blog post.
🔍 Copilot Code Reviews: The house pour-over
Good code reviews are like a pour-over coffee. They take time, care, and attention to detail. Copilot Code Reviews brings that same precision into your pull requests. It summarizes changes, highlights important details, flags security or style issues, and suggests improvements, often catching problems before your teammates even see them.
That early detection speeds up the development cycle by reducing back-and-forth on common errors. Instead of teammates spending time pointing out missing validations or overlooked issues, Copilot surfaces those issues up front. The result is a cleaner pull request that is easier for reviewers to digest.
Think of it as your friendly barista handing you a carefully brewed cup, pointing out the tasting notes, and making sure the balance is just right. Your reviewers can then focus on higher-level feedback instead of the basics. You still decide whether to merge, but the process is smoother, faster, and more enjoyable for everyone.
Learn how to request a Copilot code review or explore the code review overview.
🧠 Copilot Instructions: Be a regular
Every café has its regulars, the ones whose drinks are started before they even walk in the door. Copilot Instructions give you that same treatment. By adding a
.instructions.md
file to your repo, you teach Copilot how you and your team like things done. Variable naming, documentation tone, error handling preferences - it’s all captured.Pair it with
.prompt.md
files that describe common goals or coding styles, and suddenly Copilot feels like a teammate who already knows the house rules. No repeated explanations. No wasted cycles. Just consistent results across the whole project, every time.Check out how to configure custom instructions and read these tips for writing better ones.
🧋 Copilot CLI: The to-go window
Not every day is a sit-down café day. Sometimes you're in a hurry and need something you can take with you. That is exactly what the Copilot CLI provides. It puts Copilot directly into your terminal, so you can work faster without ever leaving the command line.
Think of it as the to-go window at the café. You step up, give a quick request in natural language, and walk away with exactly what you need. The CLI can answer questions, write and debug code, explain commands, and even interact with GitHub.com all without breaking your flow.
The real power comes from interactive mode. Start a session with the
copilot
command and you can have a back-and-forth conversation right in your terminal. You might ask it to make a change in your project, rewrite part of your README, or suggest improvements to a script. You can react to its responses and keep iterating until the task is done.It's not limited to local tasks either. You can ask the CLI to list your open pull requests, raise an issue, or even create a new branch and open a PR for you on GitHub.com. That means you can manage your project’s lifecycle without switching windows.
Learn more in the Copilot CLI docs.
🔌 MCP: Brewing with endless options
Every great café has more than just a reliable espresso machine. It has syrups, specialty roasts, alternative milks, and brewing methods that let the barista create drinks in every possible style. MCP, or Model Context Protocol, is that system for Copilot. It connects Copilot to the tools and “flavors” you need, whether that is a static analyzer, a security scanner, a custom API, or a script unique to your team.
With MCP, you're not limited to a single brew. You can call in different “ingredients” like the Playwright MCP server for automated testing or a Figma MCP server to pull in design context. Each MCP server is like a new syrup or roast you can add to the bar, giving Copilot more range to produce exactly what you want.
The best part is that the results are streamed directly into chat. Logs, diagrams, and structured outputs arrive in context, so you can see what was run, why it was used, and what came out of it. You get transparency and control while Copilot delivers the variety your project requires.
Start exploring with the MCP docs and the GitHub MCP server guide.
👥 Copilot Spaces: The team’s long table in the café
Spaces are where the café becomes more than just a stop for a single cup. It’s where the whole team lounges around a long table, sharing ideas, context, and tools. Copilot Spaces bring that energy into your code editor.
With Spaces, you ground Copilot in curated context: code, docs, open issues, pull requests, specs, and custom instructions. Copilot stops treating questions as isolated requests and starts viewing them through the lens of your project. It becomes more like a subject matter expert built into your workflow.
Once you set up a space, it can be shared across your organization. That means teammates join the same “table” and instantly access the same context, reducing repeated explanations and onboarding friction. And the context is always fresh. Files or repositories you attach sync as they change, so your space remains evergreen and never stale.
Spaces make Copilot more than your personal assistant. They turn it into a shared collaborator. When you work in a space, your prompts are richer, your feedback more aligned, and your team moves faster.
Learn more about creating spaces and using them across your team.
⚡ GitHub Spark: Dream it. See it. Sip it.
Some ideas hit you like a sudden craving. You do not want to set up a full kitchen or spend hours brewing, you just want to try it right now. That is what GitHub Spark is for. With Spark, you can turn your idea into a small, working web app simply by describing it to Copilot. No coding experience required.
You tell Copilot what you want, it builds the prototype, and you can refine it through natural language or quick edits. Hosting, storage, and a managed runtime are all included, so your app is live and shareable instantly.
Think of Spark as the tasting counter at your favorite café. You dream up a new flavor, the barista whips up a sample, and within moments you're sipping it. Spark gives you the same power for software. You imagine it, Copilot builds it, and you can share it with others to taste-test your idea.
👋 Final Sip
The Copilot Café isn’t just about convenience. It’s about craftsmanship. Every feature is a carefully brewed shot that helps you write, review, and ship code with confidence. The tools are here, the menu is growing, and your favorite drink is on the counter before you’ve even reached the front of the line.
Now it’s your turn to take a sip and see how far Copilot can take you.
Beta Was this translation helpful? Give feedback.
All reactions