Settings

Provider setup guidance

Configure OpenAI, Anthropic, or Ollama in local backend environment variables to enable chat, document Q&A, research, and coding workflows. Diagnostics show presence/status only and never reveal API keys or passwords.

Routing Controls

Ollama

Idle

Provider Priority

Diagnostics

{
  "providerStrategy": {
    "primary": "openai",
    "secondary": "none-configured",
    "optional": "ollama",
    "note": "OpenAI primary → Anthropic secondary → Ollama optional local fallback"
  },
  "routingModes": [
    "cloud-first",
    "auto",
    "fast-local",
    "deep-local"
  ],
  "defaultRoutingMode": "cloud-first",
  "timeoutMs": 30000,
  "keepAlive": "10m",
  "providerStatus": {
    "ollama": {
      "configured": true
    },
    "openai": {
      "configured": true
    },
    "anthropic": {
      "configured": false
    }
  },
  "providerPriority": [
    "openai",
    "anthropic",
    "ollama"
  ],
  "cloudProviders": {
    "openai": {
      "configured": true,
      "model": "gpt-4.1-mini",
      "role": "primary"
    },
    "anthropic": {
      "configured": false,
      "model": "claude-3-5-haiku-latest",
      "role": "fallback"
    }
  },
  "ollamaConfig": {
    "optional": true,
    "role": "optional-fallback",
    "note": "Missing Ollama never breaks the app — it is skipped if unavailable"
  },
  "recommendedFastModels": [
    "llama3.2:3b",
    "qwen2.5:3b"
  ],
  "deepLocalModel": "qwen3:8b",
  "ollama": {
    "available": false,
    "warm": false,
    "lastWarmModel": null,
    "lastWarmAt": null,
    "lastWarmError": "fetch failed",
    "models": []
  }
}