Issue with Generated but non-clickable Links in Assistant Output

Hello OpenAI Community!

I hope this message finds you well.
Over the past few weeks, I have been working on an Assistant and API that accesses a model to transform pictures of rooms into a Japanese interior style. This project has been going smoothly, with successful test runs and a seamless experience for users. Specifically, the API generated upload links for users to click on and submit their images, and once processed, it provided clickable URLs to view the transformed images in ChatGPT (

However, starting this past Tuesday, I have observed an unexpected change in behavior. While the API continues to generate the correct links, these links are no longer clickable for users in ChatGPT.

I have reviewed my code, and everything seems to be in order on my end. Therefore, I’m inclined to believe that this issue may be related to changes or updates on the side of ChatGPT assistants.

In light of this, I would like to ask our community if anyone else has come across this issue recently. If so, have you been able to identify the root cause or discover a workaround? I would greatly appreciate any insights, suggestions, or shared experiences that could help shed light on this matter.


I experience the same problem, any explanation for this?

There is a similar situation here: In custom GPT, some images in Markdown format cannot be displayed irregularly, but they can be reproduced stably under the same conditions


And here: How I Solved the GPT Image Display Glitch and Should We Keep the Security Mechanism?


thank you for your insights, this is much appreciated!

Hey folks, thanks for flagging this. We have resolved a few issues around generation URL’s. There are a few known edge cases that don’t work well:

  1. Having the model construct full URL’s on the fly (you should instead pass the full URL back in an Action response)
  2. Having the model fill in URL details (same as above, you should pass in the full URL and all UTM parameters)
1 Like

I have seen similar issues in normal chat responses. (Not action related)

For example when asked for a book link I could see some "amazon… "-ish link while html part of the message was streaming but in the end the link was not clickable and the amazon mention was gone.

Same here with my GPT. If I copy a reply containing a URL with the copy icon, I can see that the Markdown is correct, and that the URL is correct. It just doesn’t render in the browser (Safari, macOS). But it does render in the mobile app. This is surely a bug?

This is not a bug.

Please read a whole two posts up for an explanation.