JetBrains announced JetBrains Air, an ecosystem of products intended to organize agentic software development across developer, team, and organizational environments, rather than confining agent use to a single development environment window. The ecosystem combines agent execution, work coordination, policy and cost monitoring, while remaining open to models, services, and agents from multiple providers.
The company says the launch of Air brings together work that began publicly about six months ago, when it started experimenting with agentic development environments and introduced JetBrains Central as an open system for control and execution. Subsequent developments included the Central CLI interface, shared context, cloud agents, task automation, governance, and AI cost controls.
Ecosystem components
JetBrains Air consists of products that are currently available and others that will be introduced gradually. These include Air in JetBrains IDEs, an experience for directing agents, organizing their work, and validating their outputs inside JetBrains environments, leveraging the company’s tools for understanding code. Air Teams, meanwhile, focuses on coordinating and automating software delivery workflows that bring together developers and independent agents.
Air Governance carries the former name JetBrains Central and focuses on organizational policies, visibility and auditing, cost management, and accountability in AI-assisted development. Support for Junie, JetBrains’ agent for professional coding, will also continue across all Air interfaces.
The ecosystem also supports the Agent Client Protocol (ACP), which aims to standardize communication between the development environment and the complete agent ecosystem, including planning, reasoning, tools, model routing, and observability. Through the ACP Registry, developers can discover compatible agents and run them from inside JetBrains environments.
Why does this announcement matter?
The central idea is that agent adoption is advancing faster than the organizational infrastructure needed to manage it. Even if an agent can produce a software change, questions remain about access to the code, where data is transferred, the scope of human review, what happened during remote operation, who approved the change, and how it was verified.
JetBrains believes the problem shifts as agents’ role expands from producing changes to understanding, verifying, and taking responsibility for them. Code that appears correct but contains a mistaken assumption or an architectural inconsistency may carry higher review, rework, security, and infrastructure costs, even if the cost of generating it has decreased.
Multi-provider openness and current limitations
JetBrains presents Air on the premise that the future of agentic development will be multi-provider, because models and agents differ in their suitability for tasks, and teams within the same organization may choose different tools. The company says that supporting multiple models and agents is not the ultimate goal; more important is providing shared context, consistent policies, a single view of cost, and a record of what happened regardless of which tool produced the change.
Nevertheless, the announcement is partly a roadmap. JetBrains explained that Air will evolve through successive releases, distinguishing what is available, what is entering the preview stage, and what remains part of the long-term direction. Future plans include expanding mobile and remote experiences, adding context from code, architecture, repositories, runtime behavior, and organizational knowledge, as well as running more tasks in response to repository events, schedules, and delivery processes.
In practice, the announcement alone does not prove that these controls will address every governance issue or that integration between different agents will work with the same depth in all cases. But it clarifies JetBrains’ shift from a primary focus on the individual developer’s workstation toward building a layer that connects developers, agents, teams, and organizations, while making understandability, verification, and accountability part of the agentic development system.