I like lens feature a lot. I also use their privacy pass feature quite a bit. Overall there are a lot of really nice features and search results as good or better than Google. DDG uses bing which is junk IMO

Kagi

You can do a reverse proxy + authelia (or other auth service). It’s still more risky than a VPN IMO, buts wayyyy better than some of the other options in this thread

Plex clients arent great, but they are better on many TVs compared to jellyfin. Also the wife is used to it, so I don’t really want to retrain

I haven’t looked into the config details too much, but I think its audible. I did see some hardcover & goodreads related code when perusing some of the docs, but didn’t really look much deeper.

this feels like a harsh characterization. Their description of AI usage is exactly what I would want in a project using AI. I haven’t looked at their actual practices, so I guess I am trusting on that side. All that said, this is similar to my workflow at work, so I am not as negative on it.

Does this decompress the files preemptively and leave them? Or is it only decompressing as a post/subreddit is accessed? Basically i am wondering what kind of storage footprint would be required to search through this

I use git and commit configs/setup/scripts/etc. to it. I at least have a road map for how to get everything back this way. Testing this can be difficult, but it really depends on what you care about really.

• Testing my kopia backups of important data? that I manually test every once n’ while.
• Testing if my ZFS setup script is 100% identical to my setup? that’s not that important, as long as I have a general idea I can figure out the gaps and improve the script for the next time around. Obviously, you can spend a lot more time ensuring scripts and what not stays consistent, but it depends on what you care about!

For a lot of my service config, git has always worked well for me and I can go back to older configs if needed. You can get super specific here and save versions in git, then have something update the versions (e.g. WUD)

Most of the books I read are docs on all the fuggin homelab devices & services…wish I had time to read books at this point.

HA integrated with homekit well. I like to tinker, but hate doing interface/dashboard work. I did find an auto dashboard for HA which has made our lives easier. My long term hope is to use a voice assisstant for the rest.

Since it sounds like youre handy already, i would really dig into the dashboard side since that will dictate how easy it is for your family to use.

Komodo handles this better. I dont use it, but i did spend a week trying it out. You can add repositories and run builds on them then use those in a stack

I think i get what you’re saying…but its a pretty bad analogy. The way i am reading this is, most linux distros are a mish mash of various packages, window managers and apps. Many of which dont work well together, but that really hasnt been the case for the last 10-15 years. Basically when ubuntu and fedora stabilized.

I have been playing around with it. Ended up buying more recently of the lifetime stuff. Tried their android app and windows/web interface. Web interface has been 100% rock solid. Android app not so much:

• Camera sync sometimes doesn’t sync all the pictures. Sometimes it doesn’t show synced pictures either, I have to manually navigate to the folder. Not sure whats up.
• Android app has a long delay on starting up sometimes.
• Poor battery optimization comparatively speaking to alternatives (nextcloud & gphotos)
• App occasionally crashes, but that’s stopped recently.

Also, just full disclosure I am on Graphene, so sometimes things are a bit different. Recently, I have started to setup their rclone connection to use it as an offsite encrypted backup of ~ half a terabyte of important shit. Its my 2nd offsite backup and I am planning to regularly test that repository (using Kopia). While I am happy with it so far (outside the android app), I am still a bit wary on how reliable they are. They have been around for a few years now, so I feel more confident, just overall being cautious until I see a public audit of their backend. The client code is all OS (supposedly, I haven’t confirmed beyond the rclone code), so you can check that if you like (which should at least confirm local encryption before transit).

I find LLMs make a lot more.mistakes when they try to distill a guide down to steps. Its not a bad summary, but they to get confused sometimes when there are forks in the guide. That said they are really good at finding guides, especially older ones that SEs tend to not place as high

I think your best bet is to pick one thing that you can get a good guide for and start from there. If you really want to learn its probably better to start with a Debian or arch setup than proxmox, but that’s really going to depend on what you really care about learning.

I know it will be an unpopular opinion but you can use perplexity or Claude to help you find useful sources online if you’re striking out on your searching. Most of the time I find they do better with more obscure issues, but those should be rare if you’re following a guide

I tried out Komodo, but gave up on it. I looked at dockge after, but opted not to try it out. I prefer the IaaC setup with my compose in a repo for versioning and rollback. And while I think you can probably combine the two, komodo was getting in the way most of the time. It centered around secrets management and generating those secrets at run time.

That said, I feel like if I expand beyond a single server I may go back to one of these tools

I use wireguard directly instead of tailscale. Not sure what router you’re using, but mikrotik support it out of the box. I am sure they are not the only ones. My phone runs on it 24/7 and has access to the rest of my services.

I haven’t setup nextcloud, so can’t give any advice on that. Immich was insanely easy to setup though.

I like navidrome, but I am not using jellyfin, so I have nothing to compare it with.

Sure supply chain attacks are a thing, but containers aren’t the issue. Any package delivery mechanism can suffer from it. Its up to you to verify those containers and/or build it yourself

This is the correct answer. There are some occasional docker containers you can use that connect to your VPN for the service you plan to use

Grapheneos is only installable on unlocked phones. Specifically pixels. There is a complete list on their site