@nebulon Yeah, it does work. What I meant was that it doesn't work for some clients, even though they are indeed connected to a network. I used to believe it was an issue on the client side, but the fact that it works on another typebot Cloudron instance for the clients is what makes me believe it comes from the server.

hm, yeah, never mind, now it works for me as well - maybe was a browser glitch

Our ci test pipeline succeeded now and the package update is out

@shrey said in Native (email) Auth doesn't seem to work out-of-box: Recommendation: have the above config set by default, particularly in the cases where the admin selects "User Management by app", when installing the app. Those variables are always set already. See https://git.cloudron.io/cloudron/typebot-app/-/blob/main/start.sh?ref_type=heads#L31 . Or maybe you are suggesting that we should put this in /app/data/env.sh so that the user can see them?

New package with 3.0.1 is out now.

@nebulon Ok, thanks

I have tried different server instances in AWS and GCP to run Cloudron with this version 1.81, because at first I suspected there might be something wrong with the server instances, but the problem remains the same, unable to load model of LLM in respectable AI block whether it's openAI or the others, here what happens, i create new bot then I put the openAI block in canvas, choose or create new credential, after that in drop down menu below credential should come out model options to choose e.g gpt 4o, gpt 4o mini and so on, but what happened was the drop down menu just freeze showing message "load" only, and errror message said that connection to open AI API is not established, this happened to other AI Blocks, not just open AI, also to Open Router, Together, Groq. But now the problem is resolved after I upgraded the 1.81 to recent update 1.82

Pushed the package update now.

Cloudron does not support installation on subpaths. We only have domain level installation support.

@joseph Thank you, I just emailed you.

Yes, I agree, i have double checked and simulate the case couple times

This might be better suited to ask in their Discord - https://typebot.io/discord

ah the new package version is already out by now.

This is very easy to solve! You don't need any S3 storage just to serve images. That's like using nuke to bomb an ant. Here is the solution: Step1: Create a repository on your github account. Step2: Upload an image (e.g. bot avatar) to your repository. Step3: Copy the link of your image, then paste to jsdelivr (select "convert from Github"). [image: 1715934680451-d1ca53cb-6e7f-468e-9abc-e05540baea9f-image-resized.png] Step4: Copy the JSdelivr link to your Typebot. [image: 1715934804835-5677227c-ec50-4db0-95f9-38de57586ecb-image-resized.png] Now, you have a world-class CDN provider serving your image FOR FREE!!! This link is permanent unless you purge it!

I do not, but you were right, It was a plugin that I had to see if any emails were related to a website. I removed it, and it now works; thank you for the assistance!

in long: how? ^^

We only update apps to the upstream versions which are marked as latest stable.

Thank You

@roseantonio If you install the latest version of the package, it lets you install without Cloudron SSO. After that, you can set up SMTP auth like above.