Foundation Model API Integration — Connect Any AI Provider

I will integrate a foundation model API into your application — covering provider SDK setup, request configuration, streaming response handling, multi-turn conversation history management, error handling with provider-specific retry logic, and a provider abstraction layer enabling future model switching without application-level changes. Foundation model APIs have provider-specific behaviours for rate limiting, error codes, streaming protocols, and message format conventions — an abstraction layer that normalises these differences across providers is what enables switching from GPT-4o to Claude or Llama without modifying application logic.

The integration covers provider SDK installation and configuration, a provider abstraction interface, streaming response implementation, conversation history management, error handling with provider-specific retry configuration, and basic usage logging.