Hello!
Currently, our workflow analyzes the image with the OpenAI node (binary data).
When we have that data, we merge it with the rest of the info and scan all the data with the “Basic” LLM Chain" agent Open AI node.
What we want to do is to be able to analyze image AND other information in one prompt, not 2 separate ones.
Is this even possible somehow?
Why are you using an LLM chain or even a agent for this. All you need to do is a simple API call. Use “message a model” or “message an assistant”. Your instructions go in system/assistant level. Then your dynamic text goes in user level, and attach the picture.
Hey @Daniel_Lamphere
I tried both of the options previously, and I could not find a way to check my incoming JSON AND binary data in the same run. Maybe you can share how you would handle it?
But, what I found under the “AI Agent” node options was "Automatically Passthrough Binary Images ". And with this, I can describe the prompt to check incoming JSON data and incoming Image as well.
Hmmm I guess with just using the built in nodes its a little harder than I thought. Must have been some updates. But you could use something like this:
You can add the different level prompts as you see fit. I probably setup a sys prompt and then put the dynamic text into the user level.
This topic was automatically closed 7 days after the last reply. New replies are no longer allowed.
