App is now signed.

@luckow you might be interested in advanced transfer learning for your devices and models. Optimizing Mind has developed a fast way to do so with their technology called FTL or Flash Transfer Learning.

@andreasdueren thanks, I didn't know about Nova

@timconsidine Thanks but I had packaged it a while ago as a test but never published it.

If I am not mistaken, we can install outside GCP. I found tutorial on YouTube. He installed sGTM on DigitalOcean But I don't know how to implement it on Cloudron environment.

@luckow Thanks for testing, that one was actually by design I just added an override for max scanned pages (15k should do it), please rescan. Also, don't forget to set a higher crawl rate on larger pages (if there's no throttle), that will improve the speed significantly

@scooke said: the forums or help have been very very weak, or actually non-existent (no one knows what went wrong and how to fix it). This is a very good and important point. Aside from mostly Just Working (and apps that are maintained and updated promptly), one of the best things about Cloudron is how amazingly good the support is, both from staff and from the rest of the users here on the forum.

@robi Still shared plugins, but building on your idea... what if EmDash.mycloudron.com could access EmDashPlugin1.mycloudron.com, EmDashPlugin2.mycloudron.com as isolated Workers? If installing a Worker was as easy as perhaps a LAMP app, we might get isolation. Would need some header finagling to satisfy CORS. It might also be interesting to see if Cloudron could expose a "master" app like EmDash on the front-end, and access the plugin endpoints as API-like endpoints on the back-end.

Which filament did you use for the blue?

Sorry @nebulon, found it on Github. Thinking now it's more of a discussion topic, since we don't know which way you guys are leaning on Multi..

cross-linking: https://forum.cloudron.io/topic/15249/ai-devops-opencode-alternative-to-_claw-bots

New update making it even simpler to code for free. https://forum.cloudron.io/topic/15277/ai-devops-opencode-make-coding-free-with-zen-models-no-claude-openai-google

Thanks @joseph!

@humptydumpty This is a big problem. I was shocked to see that as of 2026, only 4 out of 50 states had passed the "right to repair". But if Chinese AI companies can used "distillation" to create new AI models, then perhaps the same approach could be used to create a digital twin of a tractor and then replace proprietary software with open source software.

@timconsidine what would be the right things to do from your perspective? The recently published Warm Homes Plan and Local Power Plan are both pretty decent IMHO (although still LOTS of room for improvement and not enough details about implementation yet). IMHO there is much MUCH more political corruption related to keeping us all hooked on fossil fuels than there likely ever will be with renewables (which by their very natures are inherently more distributed).

@humptydumpty Nope, not kidding. You can if you boycot the business packaging that includes the bloat. When I used to game long ago, the games were way more stable and space efficient from 3rd party repacks who ripped out all the bloat making it playable on older hardware too.

@humptydumpty Googling this didn't return any definite answers other than this post and that generic AI response which suggests using N8N and webhooks to monitor such changes and then apply different labels. https://community.vikunja.io/t/auto-tagging-for-projects/1933

@robi good approach, but I suspect your neuro-plasticity far exceeds mine !