Hey ​@nroshak,

I personally do not use this feature to be honest! Sorry I cannot help.
Did you try reaching out to support? support@airtable.com

Otherwise, you might also want to assess having this workflow run outside of airtable (e.g. Zapier, Make, or n8n to push the prompt to e.g. OpenAI and get the image back to Airtable).

Did you test how long the same image generation would last on ChatGPTs UI for instance? To understand how bad this lag really is.

Feel free to grab a slot using this link if you’d like to further discuss the 3rd party automation mentioned above! I’d be happy to show you around on a 10 min call :D 

Mike, Consultant @ Automatic Nation 
YouTube Channel

​@nroshak 

This is not yet a publicly-released feature, so maybe its speed will improve in the future.

As ​@Mike_AutomaticN mentioned above, your best bet might be to bypass Airtable’s AI and automatically generate the images using Make’s OpenAI DALL-E integrations and Make’s Airtable integrations.

I just tried this myself right now, and a very complex image was created for me in less than 8 seconds, and the image was automatically put into the Airtable attachment field for me.

Note that you will need to pay OpenAI for the privilege of doing this.

If you’ve never used Make before, I’ve assembled a bunch of Make training resources in this thread. For example, here is one of the ways that you could instantly trigger a Make automation from Airtable.

I also give live demonstrations of how to use Make in many of my Airtable podcast appearances. For example, in this video, I show how to work with Airtable arrays in Make.

Hope this helps!

If you’d like to hire the best Airtable consultant to help you with anything Airtable-related, please feel free to contact me through my website: Airtable consultant — ScottWorld

Yeah, I could use Zapier. The image prompts I’m using are very detailed infographic prompts, and it takes a couple minutes just to generate the image using OpenAI.

However, it seems quite a bit longer in Airtable, even though Airtable is using an OpenAI imagegen model.

Curious whether anyone else has seen the same. 

​@nroshak: I’m Ayesha and I work on our product team here at Airtable. I appreciate you flagging this issue, really want to make sure we get this fixed! If you could send me some more details about what you’re running into (ayesha.bose@airtable.com), we can take a look. In particular, I’d love to understand if you’re using the Automation action or the Field agent and if you’d be comfortable sharing the prompt, so we can reproduce the issue on our end. Thanks! 

Thanks ​@Ayesha_Bose , will email you - Natalka

​@nroshak I’m also seeing slow image generation. However, I cannot say for sure if it’s Airtable or OpenAI related, since I’ve done more image generation in Midjourney (in the past) and Google AI Studio (more recently).

Zapier does provide their own image generation action along with Google AI Studio support. I tried both yesterday while trying to work around the slow image generation in Airtable. But these actions also presented some other issues, including not working at times. Hopefully, all of this will get ironed out soon as companies improve things.

​@Ayesha_Bose Thanks for jumping into the conversation!

I’ve noticed slower image generation in Airtable too, so I built a small app to track some metrics:
https://airtable.com/apphxM2l3gLtDTIux/shrl3u7uXnHXaEBCt

For context, I tested the same prompts in Google AI Studio (Nano Banana) and found it to be roughly 10x faster, with surprisingly strong image quality. I’m hoping Airtable has plans to integrate Google somewhere down the line.

I’m also curious what Airtable considers an “acceptable” turnaround time for image generation. In one test I shared on LinkedIn (a gardening app), the process felt especially slow—images were being generated in parallel across multiple records, which might have contributed to the delay?

Happy to provide more details if useful. I’d love to see Airtable push image generation to the next level!