VS Code, a programmer's ally, has evolved! What are the new features that allow you to get closer to AI?
Hello everyone! Welcome to this blog, which explains AI technology in an easy-to-understand way even for beginners. I'm John, your navigator.
Today, I would like to introduce "Visual Studio CodeI'll be talking about the latest update (version 1.101, the May 2025 version!) of VS Code, a software that's like a highly functional notepad.
VS Code has a lot of great new features to help you get closer to AI (artificial intelligence). Even if you think "AI seems difficult...", after reading this article, you'll be excited to see what you can do! Let's take a look together!
What does it mean to "become friends with AI"? The secret is in "MCP"!
The most noteworthy feature of this VS Code update is "MCP (Model Context Protocol)" Our response to this technology has been greatly improved.
If you're wondering, "What is MCP? It sounds like a secret spell...", don't worry! MCP is not difficult at all.
For example, MCP is "A common language for smooth conversation between AI and apps"It's something like that.
In order for the "smart AI" we hear about so often these days (AI that can write sentences, answer questions, and come up with programming ideas) to properly understand the context of the apps we're using and provide appropriate assistance, it needs to know things like "what are we talking about now?" and "what information is needed?"Context" is very important.
MCP defines the "conventions" and "rules" to convey this "context" to AI in an easy-to-understand and efficient way. Thanks to this, various applications can pass information to AI in the same way, so AI can work smarter in more places. With VS Code's strong support for MCP, programmers can use the power of AI to work more easily and creatively.
How has MCP support evolved in VS Code 1.101? Let's take a look!
So, let me explain in detail how this AI integration function using MCP has become more convenient with this VS Code update (version 1.101)!
1. It's now easier to issue prompts to AI!
A "prompt" is a request or instruction given to an AI. For example, when asking an AI to come up with a recipe, a prompt would be to ask it to "give me three easy Japanese recipes that use chicken and can be made in 30 minutes."
In this VS Code, from the chat screen,Slash Commands” (This is a command that you use by typing “/” on the keyboard. For example, /explain (If you input ", AI will explain the selected code.) It has become easier to ask AI things like "What does this part of this program mean?" or "Can you give me an example of code to create this function?". It's like using it to give quick instructions to an excellent assistant that is always by your side.
2. Handling the "information (resources)" that AI uses as reference has become extremely convenient!
In order for AI to work smartly and accurately, it needs to refer to various "information (resources)." For example, when you ask AI to come up with ideas for new functions, related existing code and design documents are "resources."
With this update, you can now view the information used by the AI through MCP (for example, the contents of a specific file in the project, error messages, etc.) on the chat screen, save it for later reference, and even easily pass it on to the AI, saying, "Please use this information as a reference and make better suggestions!" This allows the AI to provide more contextual and more accurate advice and code suggestions.
3. MCP server can now ask AI to "try something" (Experimental!)
This may be a bit technical, but a feature called "sampling" is also experimentally supported. This means that the MCP server (a computer that passes information to the AI and receives instructions from the AI, acting as a bridge between the AI and the app) can send requests back to the AI such as "Can you try out a few patterns under these conditions?" This has the potential to make interaction with the AI more flexible and interactive.
4. Security is also assured! Supports MCP servers that require authentication
Some companies and services use MCP servers that require user authentication (verifying identity with ID and password, etc.) to protect important information. VS Code now fully supports MCP servers that require such authentication. This means that you can use the AI integration function in a safer environment, so programmers can use the power of AI with peace of mind.
5. For developers: Easier checking of MCP servers
If you are a developer who wants to make your own MCP server and make AI integration more convenient, we have good news for you. By enabling development mode, the task of "debugging" your MCP server to check whether it is working as expected has become easier than before. This will likely speed up the development of new AI integration tools.
6. Join the MCP server with VS Code's "Extensions"!
VS Code has a great mechanism called "extensions" that allows users to add convenient functions according to their preferences. With this update, extension creators can now embed MCP servers in their extensions, or automatically find and connect to MCP servers in other locations. This will expand the possibilities for AI integration, such as AI assistant functions specialized for specific programming languages and frameworks!
And more! Great new features in VS Code 1.101
Besides the MCP-related enhancements, VS Code 1.101 also includes a number of new features and improvements that will make programmers' lives just that little bit easier. Let's take a look at a few of them!
- You can now create your own "toolbox" of AI tools to use in chat!
You can now group and define the "tools" (types of things the AI can do, such as "tools to generate code" or "tools to find bugs") that you use when interacting with AI in chat format as "tool sets." This allows you to quickly enable or disable only the tools related to a specific task, such as a "tool set for web design work" or a "tool set for data analysis," making it possible to use AI more efficiently. - Program change history is now easier to see!
There is a screen called "Source Control Graph" that shows the history of when, by whom, and how a program was changed. If you select a specific change from the past, the files related to that change will be displayed immediately. This makes it much easier to track the history of a program. - Code completion is available even on the Python trial execution screen!
The programming language Python can be used in a terminal (a screen where you type commands) as you try things out while you run it, known as a "REPL." Even during a REPL session, you can now use a convenient feature called "language server completion," which suggests possible code to type next while you're typing. This is a great feature for people who are currently learning Python! - Detailed rule settings for executing program "tasks"!
A program may execute various processes (tasks), and when many of these tasks try to run simultaneously, you can now specify in detail how they should behave (for example, whether to make the new task wait or stop the old task) using a setting called "InstancePolicy".
A word from the author, John
I feel that this update to VS Code, especially the enhancements related to MCP, is a big step towards making AI more familiar to programmers and becoming a reliable "partner."
AI may seem like something far away or difficult to understand. But as the tools we use every day continue to evolve, it's great to see that even non-experts can easily take advantage of the amazing power of AI. I'm excited to continue watching how AI technology will enrich our lives and work!
This article is based on the following original articles and is summarized from the author's perspective:
Visual Studio Code boosts MCP support