Mark L.

Thanks. I had a suspicion it might be that one – the green circle / orange triange / red square icons indicate much more variance in the confidence level.

Can you tell me a bit more what a "dud" response looks like. Is is just a low-effort "Unknown", where a later run has a more firm conclusion? Or indication of error?

Hey there, thanks for the table link. I'm seeing several Claygent columns there, all using OpenAI models, private key.

Did you see this "failure on first run" issue in all three? And did this seem to affect just about every cell, or only a few oddballs?

I'd be really curious to know any specific rows/records that took a few iterations to get a suitable result.

Alright, we've resolved the the issue on our status page here.

https://status.clay.com/incidents/6e4a1a35-95db-4465-938b-d4c5484ea6b4

From our tests, editing and saving the column (no changes needed) helps make sure the fixes take effect.

Could you re-save and re-run your column, and let me know how it goes? Please let me know if you have trouble afterward.

Giorgi, we've shared this with our engineers, and added this to our status page here:

https://status.clay.com/incidents/6e4a1a35-95db-4465-938b-d4c5484ea6b4

Looks like we've identified the issue, it's just a matter of getting the fix out. We'll post updates on that page as we learn more.

Giorgi, we've shared this with our engineers, and added this to our status page here:

https://status.clay.com/incidents/6e4a1a35-95db-4465-938b-d4c5484ea6b4

Looks like we've identified the issue, it's just a matter of getting the fix out. We'll post updates on that page as we learn more.

Thank you. What I was trying to share is that the results coming back from the HTTP API column are not unique. This cell, for example, has two sets of duplicates in the list of 50 results:

So when sending a row for each item in the list of 50...

... each of those results gets a row – that cell sent 4 rows for those 2 unique share URLs. (The send table data action will create them ~simultaneously.) I'm not sure why the HTTP API column would return those duplicates, but to address it at the root, that's where I'm curious to learn more.

Excellent! So glad to hear that. Please let us know if you hit any other roadblocks – here to help!

Hey there. Good news – we just released a fix for this. In my own testing, I was able to duplicate a workbook without issue.

Could you give it another shot, and let me know if you still have trouble?

Thanks again for sharing this. Since I was able to replicate, but not find a direct fix... I've written up some notes for our engineers to help investigate and get this repaired. I'll let you know what I learn!

Hey Asad, thanks for sharing this. I'll do some testing and let you know what I find.