A beginner’s guide to JetBrains Fleet’s 2026 remote development

The shift to remote environments

Developers move to remote workflows when VS Code performance hits a ceiling. VS Code runs on Electron, which is a web browser running a text editor. This architecture causes memory usage to scale poorly. VS Code uses 3,549 MB of RAM with a folder open, while Zed uses only 222 MB. Startup speed also lags. A clean install of VS Code takes 3.00 seconds, whereas Zed takes 0.40 seconds. Extension conflicts create friction. If you rely on VS Code Dev Containers, installation of extensions can take over 8 minutes.

I find the resource drain of Electron frustrating.

VS Code lacks deep AI integration because AI features are extensions rather than core parts of the editor. GitHub Copilot cannot perform multi-file edits from a single prompt. Developers also deal with telemetry and privacy concerns. You can disable telemetry, but the process requires multiple settings. Large files cause problems. A 100,000-line file takes 1.19 seconds to open in VS Code and causes stuttering. Zed handles the same file in 0.15 seconds. Input latency in VS Code is 12ms, while Zed provides 2ms latency.

Teams looking for better AI functionality look at Cursor or Windsurf. Cursor is a fork of VS Code that indexes your entire repository to provide codebase awareness. Windsurf uses an agent called Cascade to plan and execute multi-step coding tasks.

JetBrains Fleet remote architecture

JetBrains Fleet uses a virtualized file system to host code locally, in a container, or in the cloud. Because Fleet uses a virtualized file system, you can host your codebase on a remote server, in a container, or in the cloud to keep your code and data away from your local machine. It supports remote workspaces through JetBrains Space, which lets you create a workspace for any repository in seconds. You can install Fleet on a remote machine to use a local workstation as a thin client. This setup lets you leverage higher computing capacity and avoid running heavy tasks locally.

You can sync settings.

The smart mode feature turns the editor into an IDE. It uses an IntelliJ-based engine to provide code completion, refactoring, and auto imports. Fleet supports many languages, including Java, Kotlin, Python, Go, Rust, and JavaScript/TypeScript. You can also use built-in Git support and Docker support. This tool lets you work on code with your team. You can share terminals and debugging sessions.

The smart mode takes time. The IntelliJ engine takes 3 to 4 seconds to start.

You will find the UI minimalistic, which is easy to use if you know VS Code.

Setting up the workspace

Setting up a remote server requires specific hardware. A remote host needs 1 GB of RAM, but JetBrains recommends at least 2 GB of RAM and a 2-core CPU. You can install the Fleet server on a Linux VM or WSL. For Linux x86_64, you download the binary using a curl command. You then use the fleet launch workspace -- --auth=accept-everyone --enableSmartMode command to start the server. After that, you use the JetBrains Toolbox to install the Fleet client.

You can link the client and server using a URL from the server terminal. This URL looks like http://<ip>:<port>/?code=<token>. Once you open that link in your browser, the Fleet window opens. You see a green dot when you connect to a remote workspace. If you use Linux containers with WSL, store your code in the WSL filesystem. This prevents the slowness of accessing Windows files through /mnt/c/.

I recommend using the JetBrains Toolbox for installation.

Does the smart mode delay hinder your workflow?

Pricing for JetBrains tools varies by product. For an organization, the first year of a WebStorm license costs $149. If you use the All Products Pack, the first year costs $289 per user. Individual product licenses for companies like PyCharm Professional cost $249 for the first year.