A set of AI features in JetBrains development environments. It helps explore projects, obtain suggestions and assign larger tasks to connected agents.
AI alongside the code you're reading
JetBrains AI Assistant is built into the IDE, so you can ask questions in the context of an open file, a selected section or recent changes. This helps you understand someone else’s function, explain an error or prepare a small refactor without manually moving all the context into a separate chat.
The product includes chat and editing features plus support for coding agents. The official guide distinguishes a simple suggestion from a multi-step task affecting several files. JetBrains’ general AI page also introduces Air, but AI Assistant remains a separate feature set inside the IDE.
How to convey context
Select the smallest code section needed for the question and explain the expected behaviour. If the issue involves an adjacent module, include it explicitly. Wider context does not always help: outdated files and incidental examples may distract from the current logic.
Refactoring with a clear scope
Suppose you need to remove duplicated input validation. Ask for the places it is used first, then a change that preserves externally observable behaviour. Review calls and tests afterwards. If the agent changed too many adjacent files, establish how each change relates to the original task.
Automatically generated tests need to be read too. A test that repeats the implementation line by line can pass while sharing its bug. Test behaviour instead: valid input, boundary cases and expected errors. The IDE makes tests easier to run, but cannot decide whether the requirements are correct on the developer’s behalf.
Subscription, your own API keys and agents
JetBrains describes several activation methods: its own service, external keys and supported agents. They differ in model provider, usage costs and data-handling rules. Support for a particular setup depends on the IDE version and feature, so check compatibility first.
An organisation may manage access centrally. Do not put a personal key in the project configuration or a shared repository.
Limitations that matter
The AI response can be correct syntactically and incorrect in meaning. Check the types, dependencies, error handling and behavior of the application after the change. Plans and quotas determine the available amount of cloud work, and vendor restrictions affect regional availability.
Suitable tasks
- Code explanation in IDE
- Refactoring and fixes
- Preparing tests
- Connecting coding agents
Strengths
- Editor and project context
- Several ways to activate
- Detailed JetBrains Guides
Limitations
- Compatibility depends on IDE and version
- Costs vary according to the access model
- Test generation does not replace requirements verification
Check details
Official sources
This description is based on the developer’s material. Features, prices and availability may change; check the primary source before use.
