Computer Vision demo goes blank when I click "Location"

For a couple of weeks now, when I’ve visited the Computer Vision Demo on a desktop, I can upload a photo and add a species name, but when I click “Location,” the screen goes blank instead of loading the map. I’ve tried in Chrome and Edge and the same thing happens. Is there a different link other than

I had been meaning to make a bug report about this as well, but you beat me to it :)

Can replicate, made an issue here.

Any updates on this? I’m still getting the error.

We can fix this but… why? This was a thing we made to demo something that we later integrated across the entirety of iNat. We no longer need to demonstrate the way it works since it’s now integral to everything we do. I’d be more in favor of just taking the demo down.


This is semi-embarrassing, but…I didn’t realize it was integrated across the desktop version of iNat. I’ve continued to use the old observation form because when the new one debuted, I found the old one to be more efficient in linking my observations to Flickr. I still find it perplexing (it’s not clear to me whether I’m adding one observation or several different ones when I click multiple photos), but I’ll trial-and-error and figure it out! It’s got to be more efficient than what I’ve been doing the past few months, which is downloading my SLR photos to my phone and then using computer vision on the app for things I’m not sure about.

I feel sort of dumb. Haha.


No worries. FWIW, we have a video tutorial about how to use the Uploader on the website at, Sounds like one of the more relevant parts for you is the explanation of how to combine photos into a single observation.


Thanks for the link! The direct-upload to iNat process makes sense now. I’ve just always linked my Flickr photos, and I still can’t see how to make the computer vision work with that process, even on the newer “import from Flickr” page. Is computer vision only functional on direct uploads?

I suppose I can just add my photos to both Flickr and iNat if I’m uploading a batch where there are several IDs I’m unsure of and the computer vision would come in handy. That wouldn’t take up any more time than using the demo version did. Thanks for clarifying–you can tell I’ve been using the same workflow for a long time!

1 Like

During the upload process, yes, but once you’ve imported something from Flickr, you can always just go back to it and get vision suggestions on the obs detail page.