PyCharm 2026.3 Introduces JetBrains Air for Agentic Development
New IDE release adds multi-agent orchestration, Python 3.15 lazy imports, and FastAPI dependency visualization.
1Introducing JetBrains Air
The 2026.3 release introduces JetBrains Air, a dedicated layer for agentic development within the IDE. This feature allows developers to run, track, and verify multi-agent workflows directly. It integrates with various ACP-compatible agents, ensuring that code intelligence remains active while agents execute tasks, providing a unified environment for complex automation.
2Flexible Agent Integration
JetBrains Air supports popular models like Claude Agent, Gemini, and GitHub Copilot without requiring a specific JetBrains AI subscription. Users can connect existing API keys or subscriptions seamlessly. The platform includes Junie Lite for free, enabling immediate agentic experimentation for any user with a registered JetBrains account during their coding sessions.
3Enhanced Agent Review Tools
Developers can now review agent-generated changes alongside project tools and source code. Users can anchor feedback directly to specific lines in diff views or comment within the code to provide context. This mechanism ensures that agents receive precise guidance, helping to minimize errors before any code is officially committed or shipped.
4Subproject Workspace Isolation
PyCharm 2026.3 improves workspace management by allowing individual subprojects to run in isolated environments. Each subproject maintains its own Python console and terminal, ensuring dependencies are correctly scoped. This update simplifies development for complex repositories by enabling environment-specific command execution and automatic initialization via tools like uv, Hatch, or Poetry.
5Python 3.15 Lazy Imports
The IDE now provides native support for Python 3.15 lazy import syntax. This feature defers module loading until the moment of execution, which significantly reduces startup times for applications with heavy dependency trees. PyCharm parses these imports like standard ones, maintaining full code insight and navigation capabilities for developers adopting this syntax.
6FastAPI Dependency Visualization
Pro users gain a new visual tool for FastAPI endpoints that maps complex dependency chains. The IDE draws graphs showing how Depends() and Security() calls interact across multiple files. It identifies circular dependencies that would trigger runtime errors, while also allowing developers to trace which specific endpoints rely on particular shared dependencies.
7Advanced Pydantic Inspections
Editor support for Pydantic models has been deepened to include real-time validation of field constraints. PyCharm now checks values against parameters like min_length or gt, flagging violations directly in the editor. Additionally, field names within validators now resolve to their definitions, allowing developers to jump to the source code for faster debugging.
8Refined Environment Management
The redesigned Python interpreter widget centralizes environment management for uv, Poetry, Hatch, and conda. It automatically detects existing local environments and allows for quick recreation if necessary. This update streamlines the developer experience by providing a unified interface for switching between project environments and managing Python versions without leaving the IDE.