How VS Code’s New AI Features Transform Your Coding Workflow

01 Interface Features

After the upgrade, a chat icon appears in the VS Code sidebar; clicking it opens a chat window that requires a GitHub login to use.

The UI resembles other AI editors, offering Ask, Edit, Agent modes and model provider selection, all accessible via @ and / commands.

Users can choose from various large models, including local ones like Ollama, by entering the appropriate API key.

Local model support means developers with powerful machines can run models offline, catering to enterprise needs.

The chat supports inline conversations within files and can be opened in either the editor pane or a separate window for flexible use.

02 Symbol Tricks

VS Code adopts the same shortcut symbols as Cursor and Trae, using /, @, and # to trigger different functions.

Typing / shows a list of preset AI commands for quick coding tasks.

Typing @ opens an integrated tool library for programming assistance, web queries, and custom functions.

Typing # helps locate code structures or search the workspace, improving navigation efficiency.

Examples include #githubRepo to link with a GitHub account and fetch repository information, and /new to create a new C file named algo_c automatically.

03 MCP Configuration

VS Code includes an MCP (Multi‑Chat‑Protocol) configuration panel where users can add tools via various methods such as command, HTTP, NPM, pip, or Docker.

The panel also imports previously configured MCP servers from Cursor and Trae.

All MCP service definitions can be copied into the settings.json file and started from there.

The interface shows the current status and discovered tools, confirming successful tool calls.

Conclusion

The author notes many more features like command and prompt files remain to be explored, and anticipates future articles.

While AI‑focused editors like Cursor and Trae dominate, VS Code’s AI update aims to reclaim some developers and attract non‑programmers, though ease of use for beginners remains crucial.