Since Gemini is fully multimodal for most new models now, it would be great if we could pass binary data to the Agent even if it is NOT an image (would be helpful for things like OCR of PDFs, etc)
Since Gemini is fully multimodal for most new models now, it would be great if we could pass binary data to the Agent even if it is NOT an image (would be helpful for things like OCR of PDFs, etc)