AI Providers

BraceKit supports multiple AI providers, letting you switch between models instantly without leaving the sidebar. Each provider has its own configuration, and you can use multiple providers simultaneously.

Supported Providers

Quick Setup

Step 1: Open Settings

1. Click the Settings icon (⚙️) in the header
2. Navigate to AI Provider tab

Step 2: Select Provider

Open the provider dropdown and pick a provider — search for it or use the keyboard (↑↓ to navigate, Enter to select). The configuration fields below update to reflect the selected provider. Use the + Add button to register a custom provider.

Step 3: Enter API Key

Each provider requires an API key (except Ollama):

Step 4: Select Model

Choose a model from the dropdown or type a custom model name.

Step 5: Done

Settings are saved automatically as you type. The provider is now active.

Switching Providers

To switch between configured providers:

1. Click the provider button in the input toolbar (e.g., “OpenAI ▾”)
2. Select a different provider from the grid
3. Choose a model
4. Continue chatting — context is preserved

Note: Your conversation context carries over when switching providers. The new provider sees the same message history.

Provider Features

Reasoning / Extended Thinking

Some models can show their reasoning process:

Function Calling / Tools

Most models support tool calling for MCP and built-in tools:

Image Generation

Generate images directly in chat:

Vision (Image Input)

Send images for analysis:

API Key Security

Your API keys are:

• Stored locally in Chrome’s extension storage
• Never sent to BraceKit servers
• Only used to authenticate with the AI provider

Multiple API Keys

You can configure multiple providers simultaneously:

1. Set up OpenAI with your OpenAI key
2. Set up Anthropic with your Anthropic key
3. Set up Gemini with your Google key
4. Switch between them as needed

Each provider stores its own key independently.

Custom Endpoints

For self-hosted or proxy services, add a custom provider:

1. Click the + Add button next to the provider dropdown
2. Fill in the Name, Format, and Base URL:
   • OpenAI format — LM Studio, vLLM, OpenRouter, Azure OpenAI
   • Anthropic format — Anthropic-compatible proxies
   • Gemini format — Gemini-compatible proxies
   • Ollama format — Ollama native API
3. Click Save Provider, then enter your API key in the Configuration section

See the Custom Provider guide for details.

Troubleshooting

“No models available”

• Check your API key is valid
• Verify the API key has the right permissions
• Try typing a model name manually

“API request failed”

• Check your internet connection
• Verify the API endpoint URL is correct
• Ensure your API key has sufficient credits

“Model not responding”

• Some models (gpt-5.6-sol at high reasoning levels) take longer to respond
• Check provider status pages for outages
• Try a different model

For more help, see the Troubleshooting guide.