Inputting images via File not a URL

Hi there,

I’ve noticed that when using the get-4-vision via the OpenAI API, the only way to do so is by passing a URL of the image. However, in other situations like uploading training files for fine-tuning or editing images with DALL-E 2, OpenAI API supports direct file uploads.

I’m wondering if there are any plans to introduce a new feature that allows direct file uploads when using gpt-4-vision or if this feature is already being developed.

Thank you.

I discovered that GPT-4 Vision now supports Base64 images, which perfectly solved my problem.

1 Like