I got rid of some ugly 🐛's and decided to call it a v1 by now. (previous article about the RC1) Also made it available as a docker image for easy hosting. For myself I'm using AWS Lightsail in the same way as for the GPT Assistant - read more here with some instructions.
To host it smoothly on AWS Lightsail and not being forced to change the image itself too often, I've moved the configuration for the whitelisted apps to a public URL (there's no sensitive stuff inside). You can host this config file basically everywhere, in my case it's on a S3 Bucket, but it can also be on your github or so.
Below the table for all required env-variables you will need on AWS Lightsail: