I have a scenario that contains an array of images that need to be passed to an OpenAI Assistant for Vision. I have successfully aggregated the images using an Array Aggregator. You can see that the Array Aggregator Output contains 2 image URLs in the array:
However, in the subsequent module (Message an Assistant), you can see that the input is only displaying 1 of the image URLs.
The scenario completes in full, but the output isn’t fully correct because it’s generating a response based only on 1 image instead of 2.
This would lead me to believe that I don’t have something set correctly in the ImageURLs field.
However, if I change this to accept the full Array, as opposed to the “original_file_url” portion of the array, the scenario fails:
As an FYI, I recognize that there is an “Analyze Images (Vision)” module. I can make that work successfully, however, that module is based on the Completions API, and I specifically need this working as a Thread.
Can anyone help me identify what might be wrong? Thanks!