Skip to content

Providers Settings ​

Configure AI providers and manage API keys.

What Are Providers? ​

Providers are the AI services that power Alma's chat capabilities. You can configure multiple providers and switch between their models.

Supported Providers ​

Built-in Providers ​

ProviderDescription
OpenAIGPT-4, GPT-4o, GPT-3.5 models
AnthropicClaude 3.5, Claude 3 models
Google GeminiGemini 1.5, Gemini 2.0 models
DeepSeekDeepSeek Chat and Reasoning models
AiHubMixMulti-provider aggregator service
GitHub CopilotUse your Copilot subscription
Claude SubscriptionUse your Claude.ai subscription

Custom Providers ​

Add any OpenAI-compatible API endpoint as a custom provider.

Adding a Provider ​

Built-in Providers ​

  1. Click the Add Provider button
  2. Select a provider from the list
  3. Enter your API key
  4. Click Save

Custom Providers ​

  1. Click Add Provider → Custom
  2. Enter:
    • Name: Display name for the provider
    • Base URL: API endpoint (e.g., https://api.example.com/v1)
    • API Key: Your authentication key
  3. Click Save

Managing Providers ​

Enabling/Disabling ​

Toggle the switch on any provider to enable or disable it. Disabled providers won't appear in the model selector.

Editing ​

  1. Click the Settings icon on a provider
  2. Modify configuration
  3. Click Save

Deleting ​

  1. Click the Delete icon
  2. Confirm deletion

Model Management ​

Viewing Models ​

Click on a provider to expand and see available models.

Refreshing Models ​

Click the Refresh button to fetch the latest model list from the provider.

Model Capabilities ​

Models display capability badges:

  • Vision: Can process images
  • Tools: Supports function calling
  • Reasoning: Extended thinking capabilities
  • Context: Token context window size

Enabling/Disabling Models ​

Toggle individual models to control which appear in the model selector.

Provider-Specific Settings ​

OpenAI ​

  • Standard API key authentication
  • Supports all GPT models

Anthropic ​

  • API key from console.anthropic.com
  • Supports Claude models

Google Gemini ​

  • API key from Google AI Studio
  • Supports Gemini models

GitHub Copilot ​

  • Uses your existing Copilot subscription
  • Requires GitHub authentication
  • No separate API key needed

Claude Subscription ​

  • Uses your Claude.ai subscription
  • Session-based authentication
  • No separate API key needed

Custom Providers ​

  • Base URL: Full URL including /v1 if needed
  • Use Max Completion Tokens: Enable for providers that require max_completion_tokens instead of max_tokens

Tips ​

  • Keep API keys secure - never share them
  • Disable unused providers to declutter the model selector
  • Use the refresh button after provider API updates
  • Test providers with the tool model test feature in General Settings