Am I crazy or missing something or is using image gen as an agent tool in agent builder not possible? Trying to build a simple (to me) custom label generator by passing in things like company names and logo urls. But doesn’t seem like I can use GPT Image from the agent builder easily.
I am also interested in the answer to this. I was expecting to use the gpt-image-1/gpt-image-1-nano models in agent-builder, but it’s not available.
I’m stuck with manually using https://platform.openai.com/playground/images
The official page about the API Agents (https://openai.com/agent-platform/) also lists image generation as part of the built-in tools. But that specific tool just isn’t featured in the ‘Agent Builder’ right now, as far as I can tell at least.
I also hope this feature is added as soon as possible. Without it, for some specific use cases the ‘Agent Builder’ system will be less capable than a simple ‘Custom GPT’.
I got it so for working that I get svg sourcecode, when I save it and render it, it gets a crappy image. I just want an agent to make an image for me.
Agreed, I can imagine many use cases for something like this, such as thematically consistent presentation generation.