Loading frames is cumbersome and it doesnt tel the whole story of the Video - its at best a stop gap measure for now. Why is it that if I loaded the Video via chatpgpt text interface with the file icon option, I can load a mp4 file and it does analyze it contents decently well. Why is that not available via the API?
ChatGPT is a whole product. The API is not the API for ChatGPT. The API is intended for devs to create their own products/services. Anyway, OpenAI will probably enable uploading video in the future when gpt-4o realizes its full modal potential.