Home / Integrations / Android Studio

Use TurboLLM in Android Studio

Android Studio is built on IntelliJ, so it can run JetBrains' AI Assistant — which ships a first-party OpenAI-compatible custom provider. That means you can point it straight at your local TurboLLM model with no third-party plugin required.

Before you start

Load a model in TurboLLM (or let it auto-load), then open the Developer screen and grab the Server URL and an API key — exactly the same values used in the VS Code guide.

Version requirement

The JetBrains AI Assistant provider requires Android Studio Meerkat / 2024.3.1 or newer.

Connect AI Assistant to TurboLLM

  1. Load a model and get your credentials

    In TurboLLM, load a model (or let it auto-load), then copy the Server URL and an API key from the Developer screen.

  2. Install the JetBrains AI Assistant plugin

    Go to Settings → Plugins → Marketplace, search for "JetBrains AI Assistant", install it, and restart Android Studio.

  3. Open the third-party provider settings

    Go to Settings → Tools → AI Assistant → Providers & API keys, find "Third-party AI providers", and choose "OpenAI-compatible".

  4. Fill in the endpoint and test

    Enter the endpoint URL http://localhost:6996/v1 and your TurboLLM API key, then click Test Connection.

Not seeing AI Assistant, or on an older Android Studio?

Continue also ships a JetBrains plugin (Settings → Plugins → Marketplace → "Continue") that reads the same ~/.continue/config.yaml from the Continue guide. If Continue's panel comes up blank, it needs a JCEF-enabled runtime: Find Action (Ctrl/Cmd+Shift+A) → "Choose Boot Runtime for the IDE" → pick an option labelled "with JCEF". If it still won't load, flip the ide.browser.jcef.sandbox.enable registry key (Find Action → "Registry…") to false.

Previous← VS Code