Settings UI — L.U.N.A. Docs

Overview

SettingsView (components/Settings/SettingsView.tsx) is the in-app configuration panel. It reads the current settings fromGET /api/settings/coding and saves changes viaPOST /api/settings/coding. Changes take effect on the next request — no restart required for model or provider changes.

📌

SettingsView does not manage the .env file directly — it calls the backend which updates the running settings object. To persist changes across restarts, update .env manually.

LLM provider

The provider dropdown shows all eight supported providers. The selected provider updates LLM_PROVIDER at runtime.

Option	Provider
Ollama (local)	`ollama`
Anthropic Claude	`anthropic`
Groq	`groq`
Google Gemini	`google`
OpenAI / Compatible	`openai-compatible`
Mistral AI	`mistral`
Cohere	`cohere`
NVIDIA NIM	`nvidia-nim`

After selecting a provider, a model name field appears pre-filled with the currently configured model. Change it to override the default.

Coding agent settings

The coding agent can use a different provider from the main chat — setCoding provider to "Same as chat" to inherit, or pick an independent provider (e.g. use Groq for fast chat but a local Ollama coding model for code generation).

Setting	Default	Description
Coding provider	Same as chat	Independent LLM provider for the coding agent.
Coding model	`qwen2.5-coder:7b`	Model name for coding requests.
Max iterations	10	Maximum tool-call iterations per coding request.
Shell timeout	30	Seconds before `code_run_shell` times out.

Provider status

Each provider card shows a live status indicator fetched from the backend. A green checkmark means the provider is configured and reachable; red means the API key is missing or the endpoint is down.

GET /api/settings/coding — providers field

{
  "providers": {
    "ollama":            { "label": "Ollama (local)",      "configured": true  },
    "anthropic":         { "label": "Anthropic Claude",    "configured": true  },
    "groq":              { "label": "Groq",                "configured": false },
    "google":            { "label": "Google Gemini",       "configured": false },
    "openai-compatible": { "label": "OpenAI / Compatible", "configured": false }
  }
}

Click the refresh icon to re-probe provider connectivity without leaving the settings panel.

Settings API

Endpoint	Method	Description
`/api/settings/coding`	GET	Returns current LLM and coding-agent settings + provider status.
`/api/settings/coding`	POST	Update settings. Body: partial `CodingSettings` object.

POST /api/settings/coding — example body

{
  "llm_provider": "anthropic",
  "anthropic_model": "claude-sonnet-4-6",
  "coding_provider": "ollama",
  "coding_model": "qwen2.5-coder:14b"
}