OpenAI’s current review process is stifling innovation for already approved apps.
My app has been live for over two months, yet I’m unable to push a simple bug fix due to the sluggish follow-up review funnel.
Users are suffering through poor experiences because developers can’t iterate quickly enough to address critical issues.
We desperately need an expedited review path for approved apps to ensure we can maintain quality and fix bugs in real-time.
Strongly agree with this! Similar scenario for me. Approved app needs urgent fixes, stuck in review for weeks! I’m starting to think Apps are not (or no longer) a priority for OpenAI. There are many fundamental issues (state management across turns, review processes, distribution, major bugs, etc.), which aren’t getting the attention they require.
Hey all, thanks so much for sharing!
First off, I want to make this clear: I’m not happy with the state of things right now either, between what I’ve read online and the conversations I’ve had with developers at a couple of meetups I’ve been to to attend on MCP and Apps. We as a team are looking to put more time and energy into the experience, and I think the best ammunition I can bring to the table is going to be really clear and specific stories of what’s going on, so apologies in advance if it sounds like I’m doing patient intake at the dentist. If you’d rather not share details publicly, please feel free to DM me as well.
When you said the review funnel is sluggish, when did you submit and when did you hear back? Or are you still in the pipeline?
What kind of bugs/issues are you trying to fix? Does it have to do with the tool descriptions, or the input/output tool shapes?
Could you share a bit more about the bugs you’ve run into? Even vibes off the top of your head are fine, we can try to track it down. This is currently where we’ve been jumping in in the last week, so if there are issues you’ve seen earlier, please let me know if you still see them!
I know dealing with app submissions is painful, so I genuinely appreciate that even after months of this pain you’re still willing to jump in and talk about it. I can’t promise we’ll fix everything, but we’re certainly going to do our best to make it easier to build things. Please don’t stop sharing your issues 
Thanks @casey-chow! Below some details from our experience (hope this helps with internal prioritisation).
App Review
Bugs/Issues
https://platform.openai.com/apps-manage not showing my apps for ~3 days (is working again now)
Conversation scoped state - for multi-turn app experiences (this is really important in our case)
Switching to full screen causes flicker/re-render of the widget (on desktop)
window.openai.requestDisplayMode({
mode: 'fullscreen'
});
More control over the model followup message after widget renders (href)
Why we need to update
Our app allows/requires multiple turns. We had to build our own conversation scoped state management (storing an ID). This (write action) triggers a permission dialogue for the user after the first turn and totally breaks the experience. Update focuses on a single turn and more UI focused experience.
Question: Any guidance/insight if/when apps may be invoked automatically without users having to manually add it? Discovery/distribution is effectively non-existent atm?
Hope this helps!
I had to fix a small bug within the app and have been waiting almost 2 months. The bug was realated to the UX/UI, not the tool definitions
I get e-mails from customer service reps telling me the app has passed internal checks, yet my app is still “in review”. This has happened with multiple apps, and I’m still waiting over a month since that e-mail to get an app live.
Since OAI seems to be looking for feedback (or some +1s), I will chime in too:
I’ve also seen a huge delta between when the app is reviewed (based on telemetry) and when the approval/rejection email comes. By huge I mean weeks. It’s bizzare.
I too have had everything in the app record deleted between submissions. At one point I think the validations on the icon were changed/tightened which totally broke the flow if you submitted something before that change was made.
Overall, the process is somewhere between frustrating and maddening. I’ve almost quit working on these a few times because of how long and opaque the review process.
I have two apps in review right now, both 3+ weeks and counting for this version.
Sorry I’m a bit slow here, been looking at some of the bugs surfaced. I’ll try to post updates as they come for these issues.
This was a regression and should be fixed now; if you run into this again please let us know.
A fix for this should have been just deployed as well!
@casey-chow good afternoon. Any updates on an expedited review process for already approved apps? I have a lot of small changes I need to make (e.g. mobile ux/ui optimization), but I am hesitant to spend time if it’s going to take another 90 days to get approvals.
Thanks,
Jonathan
+1 on this - we need this now. Re-approval should not take weeks, let alone months.
We are able to deploy fixes and features of our Apps much quicker now while getting feedback from users, yet the review process is the only blocker…
One other question: when resubmitting an already approved app with a new version, does the previous version stay live in the ChatGPT app directory? If so, how does that live version not pick up the latest version’s pushed changes? How does OpenAI lock / cache the entire MCP if it is just a deployment URL we are providing on submission?
Yes, I do have an update! Things should be looking a lot better and publication should be happening a lot faster now; if you are seeing response times of more than a week please open a new topic to discuss this, and I’ll see what’s going on.
The basic story here is that we had to do a pretty major infrastructural shift to scale up for the volume of app submissions we’re now publishing, so to reduce deployment risks we significantly slowed down app publication. I agree we should have been more proactive about communicating this change and I apologize for the silence on that. With hope, this is not something we will run into again.
ChatGPT apps are a very helpful addition to ChatGPT.
However, the review process needs major improvements. My app has been approved, but due to a bug, I decided not to publish it until the bug is fixed, to ensure the highest service to the users. Now, I need to wait for the 3 months again to get the fixed version approved, effectively neglecting the first approval.
Are there any estimates for the review time for already approved Apps? Should I expect the same timeline it took to get the current version approved?
