JetBrains IDEs Integration

JetBrains IDEs with AI Assistant and coding agents that work in the editor and on multi-step tasks.

Start local setup ↓Check model fit

Beginner to intermediateSetup

PartialLocal runtime

2Related paths

Recommended path

JetBrains IDEs local AI quick start

1. Start Ollama or LM Studio with a coding modelLoad a code-aware instruction model and verify its local server before changing the IDE, so connection failures stay separate from model failures.
2. Add and test the local providerOpen Settings, Tools, AI Assistant, Providers and API keys, choose Ollama or LM Studio, enter the server URL, and use Test Connection.
3. Assign models per featureBind the local model under Models Assignment for chat and core features, then configure AI Completion separately if you also need local inline or next-edit suggestions.

Open official setup docs Check what fits your hardware

Best for

Developers who want AI chat, inline help, and multi-file agent workflows inside JetBrains tools.

Model support

AI Assistant can use third-party local providers: under Settings, Tools, AI Assistant, Providers and API keys, choose Ollama or LM Studio, enter the server URL and use Test Connection. Models are then bound to individual features in the Models Assignment section.

Choose models by role

Chat and core assistance

Qwen3 Coder

Use a code-aware instruction model for explanations, refactors, and chat. JetBrains gives local providers a 64k default context, so confirm the model and hardware can sustain it.

Open the model profile →

Inline and next-edit completion

FIM or edit-prediction model

Completion is a separate assignment. Choose a provider model designed for fill-in-the-middle completion or edit prediction rather than a general chat model.

How to use this integration

Use a local model for in-IDE chat and explanation while keeping proprietary source on the machine
Assign different models to different features, so heavy reasoning runs locally and completion does not
Evaluate Ollama and LM Studio side by side against the same codebase from one settings screen
Work with AI assistance on an air-gapped or client-restricted project

Connecting a local model

Add Ollama or LM Studio as a third-party provider, verify with Test Connection, then assign the model per feature. Code completion is configured separately in the AI Completion section and needs an OpenAI-compatible provider to move off the JetBrains default.

Tradeoffs

Two documented gaps matter in practice. AI Assistant does not support invoking tools from configured MCP servers when using local models, so agentic workflows fall back to JetBrains models. And completion plus next-edit suggestions rely on JetBrains models by default, so a purely local setup covers chat far better than it covers typing. JetBrains recommends keeping an AI service subscription alongside local models to fill those gaps.

Source

JetBrains custom-model docs