Python Developer for LLM API Integration

I have a small Python backend and need help integrating an LLM API. The endpoint should accept text, send it to the model with a predefined prompt, and return the result in a consistent JSON format. The backend is built with FastAPI and I'd like to keep the implementation simple and straightforward. I'll also need basic error handling for failed API calls and unexpected model responses. This is a small, focused backend task, not a full application build. Experience with Python, FastAPI and either the OpenAI or Claude API would be ideal.