You have to edit the base_url option in configuration.yaml to the external URL of your HA instance. In my case I set this:
http:
base_url: https://hass.xxxxxxxxx.duckdns.org/
You may also have to specify the port number (default 8123) if you’re not using a reverse proxy like in my case. The it would look similar to this:
http:
base_url: https://xxxxxxx.duckdns.org:8123
I hope you get it up again!
Best
Magnus
I’ve not yet ssl configurated on my HA…
I don’t know yet how to configure it with my OVH domain 
You don’t have any access from the outside world?
Of course but via VPN
I am no expert on this particular matter, but I suspect that external access is required to verify the authentication… :-/
External access is not required.
I just verified this on my dev instance, which has neither ssl nor external access. If after you accepted the account link at the netatmo page and get an empty page just go to your HA tab and it should be authenticated.
I don’t see anymore the netatmo page but only this code:
Missing code or state parameter in http://192.168.1.53:8123/auth/external/callback?error=invalid_client&state=eyJ0eXAiOiJKV1QiLCJhbGciOiJIUzI1NiJ9.eyJmbG93X2lkIjoiMGNjNTI5NmQ2ZDBmNDBjZGE4NWU5YWI1NTNkY2MzZjUifQ.HpcWbYYpDbOOG3qctJQb6fY5sIHWvAKw5sBqrmJf6VA
In that case, the fix I proposed should get you across the finish line! 
Problem solved!
Value for
client_id: must be API from netatmo and not username
and
client_secret: must be code delivered by netatmo and not the password from the app
Thx for your help!!!
This is exactly what I did - got blank page and then returned to HA and all was goo din the hood.
Upgraded and now it’s not working, I renamed client_id and client_secret but still nothing
This caused my some serious issues, of course it’s not directly related, but as I mentioned above I want to restore to my backup.
So I flashed my sd card with the latest hassio image (done this process before several times and usually it works fine), installed samba to get my backup on the pi and restored (wipe & restore). I waited for over an hour but my ha doesn’t show up again.
SSH connection refused. It’s online because I can ping it but nothing else. Tried that twice.
I noticed that my database is pretty big (4GB). I’m running a pi4. Do I have to give it even more time? Any suggestions. I might have to open another topic for that. Maybe you can split that here @cgtobi ? Topic name could be: HA 0.105 restoring to 0.104.3 fails.
EDIT: Just now I tried ssh again and it is online again, I might give it more time. I will definitely switch to a nuc soon.
Ok, i got mine sorted, here’s how in case anyone is stuck. I commented out my entry in my configuration.yaml, restarted HA, added the integration and selected using configuration.yaml and it suceeded. then uncommented my configuration.yaml which worked on a restart
Have you removed the username and password from the configuration and have you tried to add it in integrations in the GUI?
Well, I am totally stuck trying to get my Netatmo Welcome working again. I have tried with my netatmo entry in configuration.yaml set to the three lines suggested (with my values inserted, of course), and with all the netatmo entry commented out. In both cases I can select Netatmo on the Integrations GUI. This takes me to the Netatmo site, I agree to HA access, and then - nothing! The HA GUI still displays the box asking me to OPEN WEBSITE. I have tried this with different browsers, but with the same outcome. Any suggestions?
Hi all
are Netatmo Public Data sensors still available?
I have followed the new procedure to get Netatmo climate thermostat and it works, but I don’t see anymore those sensors.
Thanks
PS: I’ve just seen that "Public weather sensors are removed and will be added back in a separate PR."
105 looks awesome but also have the netatmo problem - I am trying the same but don’t quite understand where you get the option to “select configuration.yaml”- When I go to the integration page and click +, I find the netamo integration and then it opens the oAuth page on Netatmo page where I get the Accept Loop problem.
Where do you select configuration,yaml option ?
thanks
I also get the page to authorize the third party application, wenn I press accept it closes the tab and reopens the same page again. Wenn I press authorize again, it gives me a 500 internal server error.
Same issue here with the oAuth autorization page on Netatmo. Tried creating new app with same result. On accept I get redirected and recieving an connection refused on my duckdns ssl exposed instance over port 9600. Base URL is set to the correct value, has been working for 2 years now without issue.
Also tried using the nabucase route, no luck either. Got in a loop there; after accepting page reloads after a couple of seconds.
Anybody else any pointers on where to look next?
UDPATE: it appeared there was still an older port in my baseurl (dev instance), updated that to the correct one. Also enabled popups because I saw something blocked after accepting. Result, same issue as @arjenvr , internal server error.
If you comment out all entries in your configuration.yaml and restart it is an option to select when adding the new integration before the Netatmo auth accept page, at least it was for me