Simon Willison’s Weblog

Subscribe

Running prompts against images and PDFs with Google Gemini. New TIL. I've been experimenting with the Google Gemini APIs for running prompts against images and PDFs (in preparation for finally adding multi-modal support to LLM) - here are my notes on how to send images or PDF files to their API using curl and the base64 -i macOS command.

I figured out the curl incantation first and then got Claude to build me a Bash script that I can execute like this:

prompt-gemini 'extract text' example-handwriting.jpg

Animated terminal demo. At the top of the screen is a example-handwriting.jpg with some rough handwriting. I run this command in a terminal: prompt-gemini 'extract text' example-handwriting.jpg It returns JSON showing 270 tokens used by gemini-1.5-flash-8b. Then I run the command again with -r on the end and it returns the text from the image: Example handwriting Let's try this out

Playing with this is really fun. The Gemini models charge less than 1/10th of a cent per image, so it's really inexpensive to try them out.

Posted 23rd October 2024 at 6:25 pm

Recent articles