JetBrains Fleet represented an ambitious experiment: could JetBrains build a lightweight editor that started as fast as VS Code but could activate full IntelliJ-grade code intelligence on demand? The answer was technically yes, but the market did not reward the effort. After a public preview period, JetBrains announced in 2025 that Fleet would be discontinued to focus resources on integrating AI capabilities into their established IDE products — IntelliJ IDEA, PyCharm, WebStorm, and the rest of the family.
The core innovation was the dual-mode architecture. Fleet launched in a lightweight mode that provided fast editing, basic syntax highlighting, and file navigation — competitive with VS Code for quick file editing. Activating Smart Mode spun up the full JetBrains code analysis engine, providing the deep code intelligence, refactoring capabilities, and error detection that JetBrains IDEs are known for. The transition between modes was the technology challenge that Fleet aimed to solve.
Collaborative editing was built in from the start, enabling real-time multi-user editing similar to Google Docs. This positioned Fleet for remote team workflows where multiple developers work on the same codebase simultaneously. The collaboration features were technically solid and ahead of what most IDEs offered at the time.
The polyglot architecture supported multiple languages in a single workspace without needing separate IDE products — a departure from JetBrains' traditional model of language-specific IDEs. This was appealing for full-stack developers who work across Python, JavaScript, Go, and other languages within the same project.
The discontinuation was a strategic decision rather than a technical failure. JetBrains concluded that investing AI capabilities into their existing IDEs — which already have millions of users, mature ecosystems, and deep language support — would deliver more value than building a new editor from scratch. The AI features that were developed for Fleet, including code generation and intelligent completion, were redirected to JetBrains AI Assistant across the established IDE lineup.
For developers who tried Fleet during the preview, the experience was promising but incomplete. Performance was good in lightweight mode, and Smart Mode delivered the expected JetBrains code intelligence. The plugin ecosystem was limited compared to VS Code and established JetBrains IDEs, and some language support was still maturing. The collaborative features worked well but could not overcome the ecosystem gap.
The lesson from Fleet is instructive for the broader IDE market. Building a new editor requires not just good technology but an ecosystem of plugins, community support, and workflow integration that takes years to develop. VS Code's dominance is not just about the editor — it is about the thousands of extensions, the familiarity, and the integration with the broader development workflow. Cursor succeeded where Fleet did not partly because it forked VS Code's ecosystem rather than building from scratch.