Hi, Q. What is the use-case for the DALL-E: Generate image with OpenAI/GPT Enrichment? The only thing I can get it to do is very simple image generation that is not at all personalized to the data you have. In the Loom you can see I first generated a 'fast car' but when I tried to generate a 'fast car' with the name of the company on the side, DALL-E / Clay cannot do that. It scrambles the words. Same goes for a simple prompt of asking it to 'Create an image of a sign with the name of /firstName on it' - it cannot do that. Am I doing this wrong or is this Enrichment a bit unloved and underdeveloped?
💤 You've caught us outside of our support hours (9am-9pm EST), but don't worry - we'll be back in touch within 24 hours (often sooner!). If you haven't already, please include the URL of your table in the thread below so that we can help you as quickly as possible!
Hi Duncan, thanks for reaching out. I just watched your video and don't think the correct window was shared so I can't see the incorrect images that you created through the Clay prompt - do you mind sharing those again? I also went into your Clay table to open the url links of the images that were generated and am getting an authentication failed error. It might be because the images are only viewable if I'm logged into the corresponding GPT account. Let me know if you're seeing an authentication error on your end/are able to view the images. Thanks!
Hi Tanvi R. sorry - i shared the wrong Loom originally but I then replaced it with the correct Loom, above. Here it is again - can you see the content now? thanks https://www.loom.com/share/15ae0d7e24324a29824a2e25e464648e?sid=e349cfaa-6fa0-4fff-b85d-7de1ff652143
I am now also getting the Authentication Failed message. That is wierd. But anyway - let me ask are you able to get this Enrichment to work on any table where you reference a piece of text and it puts it in the image ? Or even point me to someone successfully using this Enrichment ? thanks
Hey Duncan, thanks for reaching out! 😊 It’s not broken—it’s just AI being AI. The generative images work best when you use a well-crafted prompt. Here’s a guide on how to write a great prompt for DALL·E: How to Write Image Prompts. Just a heads-up, DALL·E isn’t a designer tool, so it’s better for creating painting-like or scene images rather than detailed, text-based designs. If you’re looking to add text to an image, there are plenty of apps out there for that. You could even create a script using ChatGPT to do it yourself if you already have the image. Here’s how I did it: 1. Add your background image to a database (like Supabase) to get a link. 2. Use ScreenshotAPI to take a screenshot of the website. 3. Add the screenshot to your Supabase database (in a Google Sheet style). 4. Download the screenshot using Python and the Supabase API. 5. Merge the screenshot with the background image using a Python script. 6. Upload the merged image back to Supabase. It can be a bit complex, but ChatGPT can guide you through each step! There are also APIs that might do this for you directly. Let me know if you need more details! 😊
Regarding the authorization failed, where exactly are you getting that from?
I get 'authorization failed' when I click on the link. the link is in the screenshot 'image url (4)'.
I added a column with colors in it and Dall-E, using a prompt from the article you linked to, did create something while using that color as the color in it. However, not a suprise, when I asked it to add a sign with 'Welcome to /color city' it couldnt' handle it. Oh well.
Ah, I see! That makes sense since they only hold the image for a short time (about 12 hours). That’s likely the reason. Keep us posted if anything changes! 😊
Hey there - just wanted to check in here to see if you needed anything else! Feel free to reply back here if you do.
We haven't heard back from you in a bit, so we're going to go ahead and close things out here - feel free to let us know if you still need something!
