AI Provider Concept

AI Providers in Ontheia form the bridge to Large Language Models (LLMs). A provider defines how and where an API is addressed, while the associated models determine which specific AI variant can be used.

Abstraction Layer

Ontheia uses an internal abstraction layer that makes it possible to:

• Address different providers (OpenAI, Anthropic, local LLMs) uniformly.
• Configure separate endpoints for tests or proxies (e.g., Azure OpenAI).
• Centrally manage authentication details.

Hierarchy

1. Provider: The technical basis (e.g., “OpenAI Production”).
2. Model: The available units of this provider (e.g., gpt-5.6-terra, gpt-5.6-luna).
3. Assignment: Agents are not bound directly to a URL, but to a combination of provider and model.