Gameyfin
The best way to support Gameyfin is to use it, break it, fix it, and talk about it. Here's how:
* Report bugs with detailed descriptions.
* Submit PRs - especially bug fixes.
* Request new features and help shape the roadmap.
* Spread the word - tell your friends, your foes, your grandma.
The more people use and contribute to the project, the stronger it becomes.
## Conclusion: Choose Freedom
At the end of the day, software is about more than just features - it's about who controls it, and who benefits from it. I hope you'll join me in supporting FOSS, not just for Gameyfin, but for all the tools we rely on every day.
https://redd.it/1pszx42
@r_SelfHosted
FileRise update: encryption-at-rest + ClamAV + hover previews + ncdu-style storage explorer (and Pro audit logs)
https://redd.it/1psvash
@r_SelfHosted
I needed offline background removal for unlimited images, so I built my own tool
https://redd.it/1psvsu2
@r_SelfHosted
The - use the space that you can - rack finally sorted and settled
https://redd.it/1psnz83
@r_SelfHosted
I just released Listseerr - auto-request Trakt/MDBList lists to Jellyseerr/Overseerr
https://redd.it/1pslqpj
@r_SelfHosted
Looking for an idea for a graduation project for university on Self-hosting/DevOps.
P.S.: I've read the rules of this subreddit, and this question doesn't seem to violate anything. If there's a better subreddit for my question, I'd be grateful for a link to it.
Hello everyone!
I'm a student who's about to write my thesis, and I really need the community's help in choosing a topic and idea for my project.
I have no problem implementing the idea (within my capabilities), but I do have trouble coming up with the idea itself.
Requirements:
1. Implementation of a final software product (web, desktop, or mobile app), i.e., not just a 100-line script.
2. Areas of interest: DevOps, self-hosting, automation, infrastructure management.
3. Relevant (no dominant, comprehensive FREE equivalent).
I really want to create something truly useful for the community, not just a "project for university." I'm willing to invest the time to develop a high-quality, in-demand tool.
The main challenge in choosing: The thesis committee has a strict requirement. The project must be relevant. If there's already a free, popular, and fully functional alternative that solves the same problem, the work might be rejected.
My initial idea (and why I doubt it):
Initially, I was thinking of creating an advanced web interface for configuring Nginx (or perhaps another proxy, like Caddy).
But there's already Nginx Proxy Manager, Traefik, Caddy-UI, etc.
And I suspect that in serious self-hosted/DevOps scenarios, people still prefer to write configs/scripts manually or via IaC rather than rely on a web UI, which may be limited in functionality (which is what I originally wanted to address).
In your opinion, is there a truly unfilled niche or unsolved problem in managing Nginx (or similar tools) that would be worth automating or simplifying via a UI to make it relevant?
Please share your self-hosting and DevOps pain points that you feel don't have a good, free, or comprehensive solution.
What tools are you missing?
I'd be very grateful for any thoughts, guidance, or suggestions.
https://redd.it/1psgl3p
@r_SelfHosted
Self-hosted URL shortener — what are people using now?
I’m looking for a self-hosted URL shortener and figured I’d ask before reinventing the wheel.
Main things I need:
short links with random IDs (`mydomain.com/a2j34yx`)
also custom/human-readable links (mydomain.com/how-to-get-to-my-shop)
something I can run myself (VPS/Docker is fine)
I’ve looked at a few options already: Yourls, Shlink, Kutt, Polr. They all seem usable, but some of them have been around for a long time and I’m not sure which ones people actually still run in 2025.
If you’re self-hosting one:
what are you using?
would you pick the same thing again today?
anything annoying to watch out for?
https://redd.it/1ps55pj
@r_SelfHosted
deeploy 0.1 - Terminal-first deployment platform for your own VPS
https://redd.it/1ps4i8u
@r_SelfHosted
What self-hosted DNS server do you use and why?
I'm looking into setting up a self-hosted DNS server for my homelab. there's a ton of options out there like pi-hole with unbound, adguard home, technitium, powerdns, bind, etc.
what do you run? why did you pick it over the others? looking for something reliable for local resolving, maybe some ad blocking, and easy to manage.
thanks for any recommendations or experiences!
https://redd.it/1ps58jh
@r_SelfHosted
Advantages of no open ports in firewall over hole punching?
Hi there, you often read about the advantages of a certain solution as it doesnt require open ports in your firewall. Well, one way or the other the communication needs to happen and often this is done via (udp) hole punching, like tailscale or newt. So I am wondering where is the adavantage other than they randomly open ports compared to standard ports which could be part of standard scanning scripts. If someone targets you and does a full nmap scan, I guess the results would be the same in both scenarios?
https://redd.it/1ps38lb
@r_SelfHosted
Navidrome/Subsonic Client Apps Catalog
Hey everyone!
I've just launched a new Compatible Client Apps catalog on the Navidrome website. You can now browse through all the apps that work with Navidrome, with filtering by platform, features, and more: https://www.navidrome.org/apps/
If you're an app developer working on a Subsonic or OpenSubsonic compatible client, I'd love to have your app listed in the catalog. You can find instructions on how to submit or update your app entry here: https://www.navidrome.org/docs/developers/adding-apps/
Check it out and let me know what you think!
https://redd.it/1prs4bu
@r_SelfHosted
Need advice on organizing Git repos and deployment strategy for multi-node homelab (no Proxmox)
# Overview
I'm running a homelab with three micro desktops on Ubuntu Server, each hosting different services. As I've added more services, keeping everything organized, version controlled, and manageable has become increasingly difficult. Coming from a software/data engineering background, I prefer infrastructure-as-code approaches with proper version control and containerization. Proxmox doesn't seem to fit that workflow, which is why I've stuck with bare-metal Ubuntu Server.
# Current Setup
# 1) mc
*Dell Optiplex 3050 Micro: i5-7500T, 16 GB RAM, 512 GB SSD*
This was my first purchase and hosts the following services:
* Private Minecraft server with [custom bash scripts](https://github.com/philiporlando/minecraft-server-utils) to automate backups and server updates (scheduled via cron)
* A [custom Python script](https://github.com/philiporlando/mortgage_refinery) that monitors mortgage refinancing opportunities
* Portainer agent
* A PostgreSQL database for freelance data engineering work (Docker Compose)
# 2) muninn
*Dell Optiplex 7090 Micro, i5-10500T, 32 GB RAM, 2 TB SSD*
Originally purchased as a database server for freelance work, now repurposed as shared storage. Currently hosts:
* Portainer agent
* A PostgreSQL database that I used as my production environment for freelance data engineering work, containerized with Docker Compose
* NFS Server (provides storage for the arr stack on huginn)
*Note: Planning to build a dedicated NAS eventually, which will free this up for more applications.*
# 2) huginn
*Dell Optiplex 7000 Micro: i7-12700, 64 GB RAM, 1 TB SSD*
Main application server, hosting my media stack and utilities:
* bazarr
* deunhealth
* gluetun
* homepage
* jellyfin
* jellyseerr (seerr)
* lidarr
* pihole
* portainer server
* prowlarr
* qbittorrent
* radarr
* sabnzbd
* sonarr
* tracearr
* NFS client (mounts muninn's NFS share on reboot)
# What I'm struggling with:
* **Repository structure:** Monorepo vs multiple repos? Should I create a private GitHub Organization to keep multiple repos organized?
* **Configuration Management:** I've read about Ansible and NixOS, but they seem overwhelming. I'm looking for something that "just works" without requiring extensive time investment to learn and maintain.
* **Deployment workflow:** Currently using VS Code's Remote SSH to connect to each host and essentially "testing in production". I make changes directly on the servers, then commit to the main branch after verifying things work. This feels backwards and fragile.
* **Version control strategy:** I'm using Docker Compose for most services and managing them with Git. Secrets and host-specific configs are handled via environment variables. Is this the right approach, or should I be doing something different?
For the dozens of us not using Proxmox, I'd love to hear how you've organized your multi-node homelabs at this scale. What tools, workflows, or repository structures have worked well for you? Any pitfalls I should avoid?
Thanks in advance!
https://redd.it/1prn1gu
@r_SelfHosted
Docker to Podman switch story
https://bogomolov.work/blog/posts/the-actual-state-of-self-hosting-on-a-vps/
https://redd.it/1prg5pu
@r_SelfHosted
Docker-ready FFmpeg HTTP API
https://preview.redd.it/7gob8wnaid8g1.png?width=500&format=png&auto=webp&s=e6672e706e38fb14fe21addc1f7ed542842a7bb2
I’ve released a Docker\-ready HTTP API for automated video, audio, and image processing, powered by FFmpeg under the hood. You can upload media, define processing steps, and get results via S3, Base64, or HTTP streaming all in a simple API.
GitHub
Get it here:
https://github.com/Aureum-Cloud/FFmpeg-API/pkgs/container/ffmpeg-api
Whether you’re building workflows, or automations with jobs, this makes FFmpeg integration way easier. Feedback and contributions are very welcome!
https://redd.it/1prgapz
@r_SelfHosted
Selfhost Notewebapp like Obsidian
I search for a note App like Obsidian, but Selfhosted.
It should support theming, and inline LaTeX
I love Obsidian but selfhosting is more fun.
I tried many options tho
\- Overleaf
\- Silverbullet
\- Siyuan
\- notesnook
\- bookstack
\- hedgedoc
and more
but i don’t like any of them :,)
any ideas?
They should deploy in a docker container easily tho.
https://redd.it/1prcor5
@r_SelfHosted
Why Gameyfin is FOSS (and why it should matter to you)
To close the current year I wrote down some thoughts I had on FOSS vs. source-available and why I think it's an important distinction for self-hosted software (you can also read it on the [Gameyfin blog](https://gameyfin.org/blog/2025/12/22/why-gameyfin-is-foss/)):
When I started developing Gameyfin, I made a deliberate choice to release it as [Free and Open Source Software (FOSS) under the AGPLv3 license](https://github.com/gameyfin/gameyfin/blob/main/LICENSE.md). This wasn't just a technical decision - it was a statement about what I believe software should be: transparent, user-controlled, and resistant to what Cory Doctorow calls "[enshittification](https://en.wikipedia.org/wiki/Enshittification)". In this post, I want to explain why FOSS matters, especially for self-hosted tools like Gameyfin, and why I think users should be cautious about source-available alternatives.
## What's the Difference? FOSS vs. Source-Available
### FOSS:
* You have the freedom to run, study, modify, and distribute the software.
* The AGPLv3 license ensures that even if someone builds on Gameyfin, they must share their improvements with the community.
* No single entity can restrict your use or lock you into a paid tier.
### Source-Available:
* You can view the source code, but the license restricts modification, redistribution, or commercial use.
* The developer retains full control and can change the terms at any time.
## Why It Matters:
With source-available software, you're trusting a single company to act in your best interest forever. If their priorities change - if they decide to monetize more aggressively, or pivot their business - your ability to use or modify the software could be restricted overnight. FOSS, on the other hand, gives you a permanent seat at the table.
## The Risk of Enshittification
Cory Doctorow's term "[enshittification](https://en.wikipedia.org/wiki/Enshittification)" describes how platforms gradually degrade user experience in favor of profit - adding ads, paywalls, or restricting features. [GameVault's](https://gamevau.lt/) paid subscription model ([GameVault+](https://phalco.de/products/gamevault-plus/checkout)) is a classic example: what's free today might not be tomorrow. With FOSS, users can always fork the project or self-host without fear of losing access to core features. The community can step in to maintain or improve the software, even if the original maintainer's priorities change.
*Note: I picked GameVault as an example because it's a well-known alternative in the game library management space and because they personally contacted me last year (more on that below). This isn't an attack (I actually favor diversity since it leads to innovation) - it's about illustrating the risks of source-available models in general.*
## Why I chose not to join GameVault
Last year I received an invitation to join GameVault as a contributor for the web UI (back then GameVault was only available as Windows client, that has changed in the meantime). While I appreciate the offer, I declined for these reasons:
1. **Philosophical Differences**: GameVault's source-available model conflicts with my belief in FOSS principles. I want to ensure that users have full freedom to use, modify, and share the software without restrictions.
2. **Diversity of Options**: I believe in a diverse ecosystem of tools. When GameVault released their own web UI, I was actually glad to see more options for users.
3. **Freedom to Innovate**: As the sole maintainer of Gameyfin, I have the freedom to experiment and innovate without needing to align with anybody else.
## Why I Don't Accept Donations
I'm lucky to have a stable full-time job that pays well enough, so I don't need or accept donations for Gameyfin. I want to be clear: this project isn't a side hustle or a way to make money. It's something I work on because I like to try out new things. That said, there are plenty of ways you can support Gameyfin - and they're all more valuable than money.
## How You Can Support
Homebox Companion v2.0.0 released! Now with Multi-Model support (LiteLLM), UI improvements, and Demo Servers
Hi everyone,
For those unfamiliar, **Homebox** is a self-hosted inventory management system aimed at home users and homelabs, useful for tracking tools, electronics, household items, warranties, spare parts, etc. It’s lightweight, fast, and designed to be simple to run.
A few weeks ago I shared here **Homebox Companion**, an unofficial companion app for Homebox that uses AI vision to scan items into your inventory instead of manually typing everything out. The feedback from this sub was genuinely useful and pushed me to ship v2.0.0.
For anyone who missed the original post, the workflow is simple: select a location, point your phone at what you want to inventory, take a photo, and let the app detect items and generate names, descriptions, labels, etc. After reviewing and approving the results, they’re submitted to Homebox. All fields are generated via structured outputs, and you can fully control the prompt per field (instructions, examples, constraints) via the settings page, including the output language.
https://preview.redd.it/l02gh0a2qq8g1.png?width=720&format=png&auto=webp&s=43171891b149141db265ce4c30956ecf643265c6
# What’s new in v2.0.0
# 🧠 Bring your own AI (LiteLLM integration)
This was the most requested feature.
Homebox Companion now integrates **LiteLLM**, so you’re no longer tied to OpenAI.
You can use:
OpenAI
Other hosted providers supported by LiteLLM
Local models (as long as they support vision + structured outputs)
I haven’t personally tested local setups yet, but if LiteLLM supports the model, it should work.
# 🌐 Try it without self-hosting first
If you just want to see how it behaves before spinning up Docker, I’ve set up two demo instances:
Homebox Companion demo: https://demo.hbcompanion.duelion.com
Linked Homebox demo: [https://demo.hb.duelion.com](https://demo.hb.duelion.com)
You can scan items in the companion app and immediately see how they land in Homebox.
>Tip: detection requests are sent in parallel, so scanning multiple items feels much faster than just one.
# ✨ Other improvements
Mostly QOL and stability work:
Fixed random logouts (automatic token refresh)
Improved mobile UI
Better PWA / iOS behavior so it feels closer to a native app
# 🛠️ Tech Stack
Frontend: Svelte 5, SvelteKit, TypeScript, Tailwind CSS.
Backend: Python 3.12, FastAPI, LiteLLM (for multi-AI support).
Deployment: Docker / Docker Compose.
Management: Managed with uv.
# 🔗 Links
Repo: [https://github.com/duelion/homebox-companion](https://github.com/duelion/homebox-companion)
v2.0.0 release notes: https://github.com/duelion/homebox-companion/releases/tag/v2.0.0
I’m especially interested in hearing from anyone running this with local vision models. If something breaks or behaves unexpectedly, please let me know.
Happy to answer questions or discuss design decisions.
https://redd.it/1psx92k
@r_SelfHosted
Who’s going to self host Spotify?
https://annas-archive.li/blog/backing-up-spotify.html
https://redd.it/1pssk2n
@r_SelfHosted
Build Your Own RSS Daily Digest in Less than 5 minutes (Matcha/Github Actions/Pages)
Hi there,
Maybe this interests this community as well, since you can also host this on your own. I made a sample digest on the Go programming language just to illustrate how to use Matcha RSS generator. Optionally you can use even password protect the static html (via staticrypt project).
If interested, I put more info on the repo:
https://github.com/piqoni/go-digest
https://redd.it/1psmg3x
@r_SelfHosted
How to get a 8Wish idle power M-ATX Xeon with ECC, Quicksync, AVX512, 8 Sata ports, 2.5gig Intel NIC, 20 PCIE lanes with bifurcation.
I thought I would share my experience with building my first Truenas home server in regards to the hardware and the fantastic build. This was my first ever actual home server build after using 3 small 1L SFF boxes for selfhosting.
After spending many months of researching hardware I finally build my server a month ago and could not be happier and I wanted to share my findings.
The parts list:
* [Intel Xeon W-1390P](https://www.intel.com/content/www/us/en/products/sku/212268/intel-xeon-w1390p-processor-16m-cache-up-to-5-30-ghz/specifications.html) $120 USD Ebay
* [Gigabyte W480M Vision W](https://www.gigabyte.com/Motherboard/W480M-VISION-W-rev-10) $200 USD Ebay
* 4x 16GB DDR4 Crucial 2666MHZ U-DIMM ECC (highest capacity ECC on the QVL, others my work but I have been burnt with non QVL ram before)
The W-1390P is basically an 11900K with ECC and proper AVX512 support and when paired with the W480 chipset has tonnes of connectivity options.
There is a lot of don't buy a Xeon if you care about power costs, but these chips are one of the exclusions to that statement as they are based off the desktop line of parts and support higher C states for great power savings.
Running `turbostat` (I know it's not the most accurate but I don't have a wall meter right now) I can see most cores are in C7 for 90%+ of the time with package power sitting between 3-8 watts usually.
I also am yet to see the CPU hit over 65 degree C when under all core 100% load when I imported 1 terabyte of videos and photos to Immich. I will admit I was worried it would be super hot but so far my Thermalright [Phantom Spirit 120 SE](https://www.thermalright.com/product/phantom-spirit-120-se-black/) has held up surprisingly well.
So far it has been fantastic and I just wanted to share my findings with everyone.
https://redd.it/1psh8qp
@r_SelfHosted
Security for private servers
I currently run nginx, nextcloud and pihole on rootless containers (some privileged), my only exposed port is wireguard for when I'm not home.
I wonder what are the most important steps to make my network more secure. I have some ideas but I don't know if they are worth the effort.
* Harden containers (low effort, but i heard that even hardened containers can be broken)
* Use VMs instead (any recommendation?)
* Use zero trust instead of wireguard (found complicated to setup, apparently I need an identity provider + zero trust + reverse proxy integration)
Any other ideas on how to improve security?
Why does it seem that the tools that are supposed to make self hosting simple overlook these important security considerations?
https://redd.it/1pscfm1
@r_SelfHosted
Comfyture: compose stack to host your own instance of ComfyUI
https://github.com/bigbrozer/comfyture
https://redd.it/1ps4pub
@r_SelfHosted
A diary about self hosting
https://redd.it/1ps63jt
@r_SelfHosted
Self-hosted web dashboard for monitoring and managing Docker containers?
hey everyone, looking for a good self-hosted web ui/dashboard to keep an eye on my running docker containers and manage them (start/stop/restart, view logs, maybe update stacks or images). i know portainer is popular but i've heard about alternatives like dockge, yacht, or others that might be lighter or better for compose files.
what do you use and recommend? bonus if it's easy to set up and works well on a single server. thanks!
https://redd.it/1ps43zd
@r_SelfHosted
Is this a secure way to put my photos online?
I want to publish my photos to an Immich install running on my NAS. I don't have any vlans -- didn't really understand the importance until recently -- so I want to make sure I get this right.
These are the steps that I'm following... is this enough to keep me secure? Are there any additional steps that I should do to make myself secure?
I don't want to use Tailscale or a VPN, because I want to make it easy for my extended family to access the photos. Am I setting myself up for failure?
**Domain & Cloudflare**
* Registered a domain with Cloudflare
* Cloudflare DNS record created for [photos.mydomain.com](http://photos.mydomain.com)
* DNS record is Proxied (orange cloud)
* Cloudflare SSL mode set to Full (strict)
* Cloudflare Origin Certificate created
* Origin cert files (cloudflare.crt, .key) placed on NAS
* Cloudflare DDNS container running in Docker on NAS
**UGREEN / NAS / Docker configuration**
* UGOS HTTPS port moved off 443
* “Redirect port 443 to HTTPS port” disabled
* Caddy bound to port 443
* Immich exposed on host port 80 -- I may change this to another random port
**Caddy configuration**
* Created a Caddyfile
* TLS configured using Cloudflare origin cert
**Router / Internet access**
* Port 443 forwarded from router → NAS
* Any inbound ports forwarded
* Tested access from outside LAN
**Immich sharing goal**
* Immich accessible via public HTTPS
* Family can access photos without VPN
* Public share links allow anonymous access
https://redd.it/1pruobm
@r_SelfHosted
Understanding the security of PUID/GUID environment variables in Docker
Recently, I have embarked on a journey to harden my security. I have a bare metal Debian host that runs several docker containers. For some of the containers, the image provider allows for PUID and GUID as environment variables.
My mental model of how this works is that when I spin up a container, it runs as root first, then it does whatever the image told it do. Finally, it changes all the relevant permissions to match the PUID and GUID. (please correct me if I'm wrong, I couldn't find information much about this)
Am I correct to say that by using PUID and GUID, I am trusting that the image owner doesn't do anything malicious while they root access before handing off to the specified PUID and GUID?
If so, how else can I ensure total rootless behaviour? Is the solution going to Podman (or does this cause even more issues)?
Also, I have been playing around with the user: uid:gid field in Docker and manually changing the bind mounts permissions to match them but a lot of images break due to not having root access for other reasons.
https://redd.it/1prl98v
@r_SelfHosted
Is pangolin bloated
Hi guys,
Past few weeks I am struggling to get tunnels optimize,
I removed all the resources and pangolin is idealing at 500 mb to 800 mb memory,
Tunnel speed is almost 1 to 2 mbps. ( Newt)
Same vps wiredoor is giving 85 to 100 mbps
Any help or insights. Or its time to go back to traefik+ wg .
CF tunnel crack down pointed me to move all my resources to pangolin but it too resources consuming with almost dead performance.
https://redd.it/1prl4jo
@r_SelfHosted
IoC
How many of you deploy your services via tools like Terraform/ ansible/ etc? For me, I do it for the education (I use it in my capacity of a System Admin) but, honestly, without the experience/education component it would likely be overkill. Still, I really love the idea and cleanliness of IaC.
How about you?
EDIT: Fixed typo (IoC -> IaC)
https://redd.it/1prguak
@r_SelfHosted
There is a limit to self-hosting.
Ever since I got into self-hosting, I’ve been hanging around a bunch of communities. Whenever someone shares their setup, I always read it closely, trying to find that one missing piece I didn’t know I needed.
After a while though, I realized that once your setup reaches a certain scale, most people’s needs start to converge. You don’t actually need more self-hosted services anymore. But the itch to try new stuff never really goes away, so you keep tinkering anyway.
TL;DR: I don’t really need any new self-hosted apps, but I still can’t stop looking for new things to install.
https://redd.it/1preqxc
@r_SelfHosted
One account to access my services.
https://redd.it/1pr86ad
@r_SelfHosted
