The text model is chosen by Sprwt in a superadmin setting (Sprwt AI: Choose Which Model To use). In the current version the default is Google Gemini 2.5 Flash, and the available options are GROQ (Llama), OpenAI GPT-4, GPT-4o, GPT-4o Mini, Gemini 2.5 Flash, Gemini 2.5 Pro and Gemini 2.0 Flash. If the stored value is ever invalid, the system falls back to GPT-4o Mini so AI features keep working. Image generation always uses OpenAI's image model, and receipt scanning always uses GPT-4o vision, regardless of the text model. This setting is not exposed in Client Config, so contact Sprwt support if you want a different model.