Configuration
Configure LLM providers, models, and basic settings.
First-Run Setup
When you first open Sophon, the Setup Wizard guides you through adding an LLM provider. Sophon ships with no pre-configured providers — you bring your own API keys.
Adding an LLM Provider
Via Dashboard
- Navigate to Settings > Models & Providers
- Click Add Provider
- Select your provider — Sophon supports 30+ providers (Anthropic, OpenAI, Google Gemini, Azure, Bedrock, Groq, Mistral, xAI, local Ollama/LM Studio, and more)
- Enter your API key — or use a subscription / CLI sign-in instead
- Test the connection
- Set priority and budget limits
Via CLI
sophon models addThis starts an interactive wizard that walks you through the same steps.
Supported Providers
Sophon supports 30+ model providers — cloud APIs, local runtimes, OpenAI-compatible endpoints, and subscription/CLI sign-in. A few common ones:
| Provider | Type | Auth |
|---|---|---|
| Anthropic (Claude) | Cloud | API key or Claude Pro/Max subscription |
| OpenAI (GPT) | Cloud | API key or ChatGPT subscription |
| Google Gemini | Cloud | API key |
| Azure OpenAI | Cloud | API key + endpoint |
| Groq, Mistral, DeepSeek, xAI (Grok), Perplexity, OpenRouter, … | Cloud | API key |
| Ollama, LM Studio | Local | None (base URL) |
| GitHub Copilot | Subscription | CLI / OAuth |
See the complete list with provider ids in Supported Providers, use a subscription instead of an API key with Subscription & CLI Sign-In, and learn how Sophon chooses and fails over between models in Routing, Failover & Budgets.
Configuration Files
All configuration lives in ~/.sophon/config/:
models.json— LLM provider configurations (encrypted)channels.json— Channel connections (encrypted)appsettings.user.json— User-level overrides
Next Steps
With a provider configured, create your first agent.