AI Providers¶
Stridevox supports multiple AI providers for the chat feature. You can switch providers at any time without losing your conversation history.
Supported Providers¶
| Provider | Models | Get API Key |
|---|---|---|
| xAI (Grok) | Grok-2, Grok-3 | console.x.ai |
| OpenAI | GPT-4o, GPT-4, GPT-3.5 Turbo | platform.openai.com |
| Anthropic | Claude Sonnet, Claude Haiku | console.anthropic.com |
| Google Gemini | Gemini 2.0 Flash, Gemini Pro | aistudio.google.com |
| Azure OpenAI | GPT-4o, GPT-4 (via Azure) | portal.azure.com |
Configuring a Provider¶
- Go to Settings (click the icon)
- In the AI Provider section, select your provider from the dropdown
- Enter your API key in the corresponding field
- Select a model from the model dropdown
- Click Save Settings
API keys are encrypted
Your API keys are encrypted using Fernet encryption before being stored. They are never logged or exposed in the UI after saving.
Checking Your Key¶
When an API key is already configured, the input field shows "configured " instead of the actual key. To change it, simply paste a new key and save.
Azure OpenAI (Additional Fields)¶
If you're using Azure OpenAI, you'll need to fill in three extra fields:
- Azure Endpoint URL — Your Azure OpenAI resource endpoint (e.g.,
https://your-resource.openai.azure.com) - Azure Deployment Name — The name of your model deployment
- API Key — Your Azure OpenAI API key
Choosing the Right Model¶
Recommended models for fitness analysis
| Provider | Best Model | Why |
|---|---|---|
| xAI | Grok-3 | Fast, excellent at data analysis |
| OpenAI | GPT-4o | Great all-around, good reasoning |
| Anthropic | Claude Sonnet 4 | Thorough analysis, clear explanations |
| Gemini | Gemini 2.0 Flash | Very fast, good for quick questions |
Larger models generally give more detailed and nuanced fitness insights, but are slower and more expensive per query.
Switching Providers¶
You can change your AI provider at any time:
- Select the new provider from the dropdown
- Enter the API key for the new provider
- Save Settings
Your existing chat history is preserved — switching providers only affects new messages going forward.
Cost Considerations¶
Each AI provider charges based on usage (tokens processed). Typical costs per chat message:
- Simple questions: fractions of a cent
- Detailed analysis with large context: a few cents
Check your provider's pricing page for exact rates. Stridevox sends your fitness data as context with each message, so longer chat histories use more tokens.
Managing costs
- Start new chat sessions for unrelated topics (smaller context = lower cost)
- Use faster/cheaper models for simple questions
- Use advanced models when you need detailed analysis