Model provider

Answers and contract processing use one selected provider. FATHm never switches to another provider automatically.

Current route

Loading configuration…

Prepare a provider choice

This prepares configuration for review. It does not activate a provider or send data. The choice applies to the whole deployment after an operator applies it and restarts FATHm.

Provider General OpenAI model (required fallback)

General reasoning effort

Optional task models override the general model. Blank task models use the general model. A dedicated model with blank effort uses the provider's default; it does not inherit another model's effort. Review model support before choosing an effort.

Claude tier 2 model Claude tier 3 model

Use separately approved API access and billing. A ChatGPT subscription does not supply an API key. Do not enter credentials here.

Configuration awaiting review

Demonstration mode stays on. Turning it off requires approved access, billing, retention, and permission to send the intended data.