Structured OCR with GPT Vision

Here's a direct link to the code: https://github.com/caesarnine/llm-experiments/blob/main/3_st...

Just a Pydantic model to define an extraction schema + a simple prompt works reasonably well.