Hi, Ive just created a new pickaxe. But when I test the upload photo feature it doesnt upload the the photo into the chat so the chatbot cant identify the image. Has this been solved?
Yes, you can upload photos into the chat. However you must use the GPT-4o or GPT-4o mini model. Image recognition does not work with other models currently.
Here is a tutorial video that walks through how to build chatbots with image recognition.
thought so. thanks mike. this is awesome btw! appreciate the work you guys are doing for newbie AI entrepreneurs
i’ve been trying to save the photo using make.com but can’t seem to get the photo into google sheets with the output data. I even tried going into google drive, but nothing. Is this possilbe or am I spinning my wheels?
What are you trying to save? The photo that the user uploaded? If so, I am actually unsure of how to save it. In reality what’s happening in the backend is the photo is sent to a vision model which ‘looks at’ the image and then returns a very, very detailed text description of the image to your Pickaxe. So your Pickaxe is actually looking at text and then can respond to it.
The upshot is even though we show the image in the conversation in the front-end interface (for the benefit of the end-user), the photo is not actually there. A very detailed text description of the image is actually there in the backend.