Platform Settings

AI enrichment fallback chain

When enriching a product, the system resolves the AI provider, model and API key in this order:

1. Feed source override — set per-feed in Feed Sources → Settings tab → AI Provider Override
2. Tenant key — set per-tenant in Tenants → Edit (AI Settings section)
3. Platform default — OpenAI key configured in server environment (currently active if no tenant key is set)

The AI Usage page shows which model is being used. If you see "gpt-4o-mini" without setting a model, the platform default is active.