Configuring models

Language Model Configuration

Aide comes pre-packaged with a few language model defaults that are always best-in-class, and requires zero configuration at your end to get started. However, if you do wish to use a language model of your choice, or even choose a provider you prefer, you can configure this in Aide.

Note

Do note that our prompts are currently written to deliver the best performance from the pre-configured models. Performance may vary/degrade when using custom models, and we will not be able to investigate issues where these models don’t behave as expected.

Quick Model Switching

You can quickly switch between models using:

• Keyboard shortcut: ⌘/Ctrl + K + M
• Command palette: Search for “Model Selection: Switch Model”

Recommended Models

For hosted solutions:

• Claude Sonnet 3.5 (by Anthropic) is currently the best performing model.
• For self-hosted or OSS models, we recommend Qwen models, which have shown excellent performance across various editor scenarios in our testing.

Provider Setup

Access provider settings by pressing ⌘/Ctrl + Shift + ,. You’ll see a configuration page like this:

Below are the supported providers and their setup instructions:

Anthropic

1. Get your API key from Anthropic’s documentation
2. Enter the API key in the Anthropic provider section

OpenAI

1. Obtain your API key following OpenAI’s guide
2. Optionally set a custom API base URL for private deployments

Azure OpenAI

Required configuration:

• Deployment ID
• API URL
• API Key

TogetherAI

1. Create an account and get started using their quickstart guide
2. New users receive $25 in free credits to experiment with different models

Ollama

No configuration needed - just ensure Ollama is running in the background.

LMStudio

Configure the local URL if using a custom port for the LMStudio API server.

FireworksAI

1. Generate your API key following their quickstart guide
2. Known for fast inference speeds - recommended for copilot-style models

Custom Model Configuration

For services like OpenRouter, you can customize model IDs:

1. Select any model from the model selection page
2. Rename it and update the model ID as needed

Example: Configuring Gemini-2.0-flash model:

Switching Providers for Models

You can run the same model through different providers:

1. Click the edit button next to your model
2. Choose your configured provider from the dropdown

Need help with a provider not listed here? Join our Discord community - we’re happy to add support for your custom infrastructure and help with setup.