I redesigned Calibre-Web
https://redd.it/1rgm2xu
@r_SelfHosted
Shelfmark 1.1.0 now available - Multi-user support, request and approval system, OIDC auth, notifications, new audiobook sources and more
Hello everyone, thanks so much to everyone who downloaded Shelfmark when it updated from CWA Book Downloader a couple months back. The sheer volume of people trying it and some of the support with PRs and issues has been awesome, so huge thanks for everyone who's given it a try since the update.
Repo with docs and the docker compose is here: [https://github.com/calibrain/shelfmark](https://github.com/calibrain/shelfmark)
As a quick recap of what Shelfmark does. Shelfmark is designed to be a simple user interface for searching and downloading books / audiobooks for your library, pulling in different download and metadata sources together in one place.
There's a wealth of options and features to help with this goal, a few examples among many:
* AA, Torrent, Usenet and IRC sources, with fully configurable clients and indexers
* Built-in protection bypass for web sources, no need for Flaresolverr or similar tools.
* Custom output options including directory creation and renaming templates, Booklore API upload, automatic SMTP email attachment support, torrent hardlinking, and much more. These can be set independently for books and audiobooks.
* Full networking tools including automatic DNS switching, HTTP/SOCK5 proxy config, and full Tor routing if needed.
* Search via Hardcover, Open Library, or Google Books available.
* Custom scripting for more advanced setups
[The Shelfmark UI](https://preview.redd.it/sizaolzli9mg1.png?width=3024&format=png&auto=webp&s=a838d098e1be15bd261cb2610f8dd1e9c2dd9de5)
\---
1.1.0 adds the remaining big feature set for Shelfmark:
**Multi-user support:**
* Configure users and admins with independent output and notification settings
* Works with local users, reverse proxy auth, Calibre-web users, and OIDC
**Request and approval system:**
* Configure users to request books instead of downloading
* Can be configured for all sources, or tweak per source and per content type (E.g. allow direct downloads, require requests for Prowlarr downloads, block audiobooks entirely, etc).
* Request rules for sources and content type can be also set at individual user level.
* Per-user download visibility - non-admins only see their own downloads
* Works with all auth types.
[Example of requests and downloads](https://preview.redd.it/lr91w8jtg9mg1.png?width=3024&format=png&auto=webp&s=f7a74be0421ca7a3f0569d7457eb95f596da5684)
**OIDC authentication**
* Includes auto-discovery and provision of users within Shelfmark
* Works seamlessly with the multi-user setup for custom settings and requests
**Notification support**
* Configure notifications for all request and download events
* Can be configured globally or per-user
* Uses Apprise - works with basically any notification provider
**New source - Audiobookbay**
* Adds the ability to search ABB torrents within the Shelfmark UI
**SMTP Email output mode**
* Optionally send completed book files as email attachments
* Emails can be set per-user
This largely completes the feature set for Shelfmark in terms of big additions. Focus from now on will be stability, fixes, and other small scale refinements that support the main search goal of the app. Huge thanks to everyone who's suggested ideas, reported issues and submitted PRs so far, and any further contributions to help with bugs and maintenance is always appreciated. Thanks!
[https://github.com/calibrain/shelfmark](https://github.com/calibrain/shelfmark)
https://redd.it/1rh71kg
@r_SelfHosted
Sub-SubReddit for SelfHosted
Just a question that has been going round my head for a few days.
What ever your opinion on it. We need to acknowledge that "Vibe Coding" is growing, and its not going to decrease.
The fact the entry point is so low from a skills perspective, and time required to pump them out, effectively makes them disposable.
Members here have been concerned for a while about the quality, security, longevity of these apps, and it turning into a flood - and as if on que, over the past few weeks, we've seen exactly this - We've been flooded (this place is unusable on Fridays). Now, issues are identified in the code, and rather than facing and fixing them, the "devs" are running off, shutting down the repositories etc.
This leads to 2 conflicting issues
1) There is an open hostility to those who share vibe-coded apps - where we see outright hostility and vitriol language.
2) There are going to be some quality vibe-coded apps, where they are properly developed, supported and managed.. as community, they would benefit members.
So my question is - how do we bridge this? My I've been thinking on this for a few days, and the only solution I can see, we have an associated sub (same mods etc) for these apps to be posted. maybe after a time, once they've proven they are well run and have longevity, they can "graduate"... lets call it SelfHosted-Vibe or VibeHosted.
https://redd.it/1rh31si
@r_SelfHosted
BookLore 2.0 is out! Audiobooks, multi-format books, overhauled readers, and a lot more..
BookLore 2.0 is here.
For the unfamiliar, BookLore is a self-hosted digital library for ebooks, comics, and now audiobooks, with multi-user support, smart shelves, metadata fetching, Kobo/KOReader sync, OPDS, and built-in readers.
Also, we just hit 10K stars on GitHub, which is wild. Huge thanks to everyone who's contributed code, bug reports, and feedback over the past year. If you want to support the project, you can sponsor on Open Collective or Ko-fi.
Here's what's new:
Multi-Format & Audiobook Support
A single book entry can now hold EPUB, PDF, CBZ, and audiobook formats all under one roof. No more duplicate entries for different formats.
Dedicated audiobook player with streaming playback, session tracking, narrator metadata, and sidebar filtering by narrator.
Reader Upgrades
PDF reader: annotations (highlights and notes), dark/light mode, range streaming for large files
Ebook reader: fullscreen, keyboard shortcuts, search, go-to-percentage navigation, text copy
Comic reader: fullscreen, slideshow mode, RTL reading direction, long strip mode, keyboard navigation
Fully bidirectional and user-scoped, reading progress syncs both ways
And a lot more:
10 new statistics charts (reading pace, distributions, heatmaps, completion tracking, etc.)
Dedicated series browser with search, filtering, and sorting
Duplicate book detection and merging
Shelves and magic shelves sync as Kobo tags
Annotation notebook collecting all your highlights in one place
Audit logging for admin actions
Author bios and images via Audnexus
Sidecar metadata file support (.metadata.json)
Content restrictions (age and content ratings) for magic shelves
15+ languages supported through Weblate
Upgraded to Java 25, Spring Boot 4, removed nginx (Angular served directly from Spring Boot)
Bookdrop folder polling, login rate limiting, upload progress tracking, multi-field sorting, and a long list of bug fixes
A note on the v2 launch
Apologies to those who ran into issues during the initial v2.0.0 rollout. Some setups broke due to file permission problems, especially for Portainer, Unraid, and Proxmox users. In v1.x the Docker container ran as root because the embedded nginx required it. In v2 I removed nginx to pave the way for a rootless container, but I overlooked some edge cases in that transition. With 200+ commits and just me doing all the manual regression testing, I missed some critical areas. That's on me.
Things are in a much better place now with v2.0.4. If you're upgrading and still hit permission issues, make sure your data, bookdrop, and book folder permissions match the user/group set in your Docker Compose file. You can also check GitHub Issues and Discord where others have posted fixes. I'd also really appreciate help with testing before future major releases, so if you're interested, hop over to Discord.
Try it out: demo.booklore.org (user: booklore, pass: 9HC20PGGfitvWaZ1)
If you're enjoying BookLore, a star on GitHub helps more people find it.
FAQ
Q: You use a lot of emojis and ship features fast. Is this vibe coded? No, I just like emojis 😄. BookLore is AI-assisted, not vibe coded. I've been building software for years, going back to early Android (anyone remember Cupcake, Donut, Eclair?) and late Symbian. The project started before AI tooling went mainstream, so all the architecture you see was built by hand. You're in good hands.
Q: Any plans for mobile apps? Yes! I've been prototyping iOS and Android apps for a few months and they're getting close. No ETA yet, but actively in the works.
Q: BookLore uses a lot of RAM, any way to reduce it? That's a tradeoff of running on the JVM, but I think the benefits outweigh it. Java has also been
How do I handle internal certs the most "invisible" way
I honestly just get sick of the insecure warnings, the inability to use the "copy" java button, and a host of other crap on my internal dockers. And Chromium based browsers won't even let you override this when you want.
Frankly on a LAN, all of that is just annoying.
Options
1. Reverse proxy - I don't like this because I use raw IP addresses all the time and have them memorized in my home lab.
2. Acme based certs - some dockers support this, others do not
3. Adding sidecars and other apps to help get them show their certs - seems like more crap to manage.
Honestly if I could just have my browser use a total unsecure mode on my LAN, I would. I just want things to be smoother and easier.
Ideally if I access a docker like, let's say, aiometadata at 192.168.1.101 or whatever, I want it to use https. If I access it at aiometa.home, I want a cert. I don't want to think about it, I just want it to work...
Is there a way to solve this that is relatively simple and as automatic as possible?
https://redd.it/1rbb2y6
@r_SelfHosted
Building a Solar-Powered Bird Station with BirdNET-Go
https://alec.is/posts/building-a-solar-powered-bird-station/
https://redd.it/1rb10fz
@r_SelfHosted
This how I feel, but only thing I do is copying docker-compose.yml and up -d
https://redd.it/1rawdyg
@r_SelfHosted
How to add a poison fountain to your host to punish bad bots
I got tired of bad bots crawling all over my hosts, disrespecting robots.txt. So here's a way to add a Poison Fountain to your hosts that would feed these bots garbage data, ruining their datasets.
[Apache](https://gist.github.com/jwakely/a511a5cab5eb36d088ecd1659fcee1d5)
Discourse
[Netlify](https://gist.github.com/dlford/5e0daea8ab475db1d410db8fcd5b78db)
Nginx
This is an amended version of an older reddit post
https://redd.it/1rasbox
@r_SelfHosted
Betterlytics - Self-hosted Google Analytics alternative with uptime monitoring
https://redd.it/1r9y3zf
@r_SelfHosted
Dispatch - A Local To-Do and Journaling App
https://redd.it/1ra257e
@r_SelfHosted
BrainRotGuard - I vibed-engineered a self-hosted YouTube approval system so my kid can't fall down algorithm rabbit holes anymore
https://redd.it/1r9u2x4
@r_SelfHosted
I got tired of naming my scanned documents so i built this !
https://redd.it/1ra0m17
@r_SelfHosted
MediKeep - Personal Medical Records Keeper
For some time i've been searching for a selfhosted solution to keep my medical records (lab tests, weight, blood pressure, medical visits, etc., etc) and ditch Withings online services and Excel spreadsheets.
A couple of weeks ago i've came across MediKeep after it was mentioned on the Selfh.st newsletter and i've using it since then. I can say it is just what i was looking for and much more. It's been actively developed and its dev is a very approachable person, promptly fixing bugs that are detected and still adding features.
It has:
\- Multi patient
\- Medication tracker
\- Lab results history with trend analysis and a fair large database of blood work tests
\- Treatment plans linked to medications
\- Procedures history
\- Vital signs history (blood pressure, heart rate, temperature, weight, BMI, blood glucose and A1C) with trend analysis
\- Symptoms tracker and history
\- Conditions
\- Allergies
\- Injuries tracker
\- Immunization tracker
\- Visit history
And more like insurance details, emergency contacts, practitioners, pharmacies, etc. It also has a feature to export records so you can take them with you to show to your health care professional.
If you're looking for something very complete to centralize the health information of your family in a clear and simple way, be sure to check it out:
https://github.com/afairgiant/MediKeep
Disclaimer: i'm not part of the dev team but im currently a sponsor since good foss should be rewarded.
https://redd.it/1r6auq1
@r_SelfHosted
Change my mind: There is no good alternative to Discord (yet?)
There is no alternative which offers:
(group) voice chat, (group) text chat, live screen sharing, permission system
easy selfhosting with docker
open source
respects privacy (looking at you matrix)
decentralized at best
costs nothing
Checked:
XMPP based services
Matrix
Stoat
Mattermost
https://redd.it/1r5stz9
@r_SelfHosted
Turning a physical mailbox into an event source using Zigbee, MQTT, and Home Assistant
My mailbox is in the building entrance hallway, several walls away from my flat, with no power outlets nearby.
I wanted a push notification the moment a letter arrives, not when I open the mailbox.
I used an Aqara P1 motion sensor inside the mailbox, but it couldn’t reliably reach my Zigbee coordinator because of distance and metal interference.
The solution was flashing a Sonoff Zigbee Dongle-E with router firmware and powering it from a power bank to extend the mesh closer to the mailbox.
Unexpected issue: the power bank kept shutting down because the router draws too little power. I fixed it by adding a small USB keep-alive module (dummy load) to prevent auto shutdown.
Now the flow is:
Sensor → Zigbee router → Zigbee2MQTT → Home Assistant → instant push notification
I also added a watchdog using last_seen so Home Assistant notifies me if the router goes offline.
Works reliably now.
Full breakdown with architecture diagrams and setup:
https://jonathas.com/how-i-made-my-physical-mailbox-send-me-real-time-push-notifications/
https://redd.it/1r5jznp
@r_SelfHosted
Portabase 1.4.0: OIDC Support, New OAuth Providers, and Improvements
https://github.com/Portabase/portabase
https://redd.it/1rh418c
@r_SelfHosted
Grafana dashboard to tell me how expensive my hobby is
https://redd.it/1rgh646
@r_SelfHosted
making strides in memory efficiency lately. In v2 I added a flag that should bring RAM down by roughly 10-15%.
Q: Any plans to move to PostgreSQL? Not at this time. BookLore is deeply tied to MariaDB and migrating would be extremely disruptive given the user base we have now.
Q: What are your future plans for BookLore? Short term, stabilizing the v2 release. After that, maybe some social features? I'd love to hear what you all want. Come hang out on Discord or drop suggestions in GitHub Issues.
https://preview.redd.it/hh6lb7f4m5mg1.png?width=3474&format=png&auto=webp&s=7f7e4d2a28c39baa60689f055a6ced64dc9c3265
https://preview.redd.it/6ki8je45m5mg1.png?width=784&format=png&auto=webp&s=00c37f7b5bc8a72c1763ad8a86c34f9ef9ae9f28
https://preview.redd.it/1qpsri2dn5mg1.png?width=1520&format=png&auto=webp&s=c63c982cdabf1d4b600a3e839447af5e46414761
https://redd.it/1rgrkhz
@r_SelfHosted
Large US company came after me for releasing a free open source self-hostable alternative!
TL;DR: I made an open-source, local-first dashboard for drone flight logs because the biggest corporate player in the space locks your older data behind a paywall. They found my GitHub, tracked my Reddit posts, and hit me with a legal notice for "unfair competition" and trademark infringement.
Long version: I maintain a few small open-source projects. About two weeks ago, I released a free, self-hostable tool that lets drone pilots collect, map, and analyze their flight logs locally. I didn't think much of it, just a passion project with a few hundred users.
I can’t name the company (let's call them "Company A") because their legal team is actively monitoring my Reddit account and cited my past posts in their notice. Company A is the giant in this space. Their business model goes like this:
You can upload unlimited flight logs for free.
BUT you can only view the last 100 flights.
If you want to see your older data, you have to pay a monthly subscription and a $15 "retrieval fee."
Even then, you can't bulk download your own logs. You have to click them one by one. They effectively hold your own data hostage to lock you into their ecosystem. I am not sure if they are even GDPR complaint even in the EU
To help people transition to my open-source tool, I wrote a simple web-based script that allowed users to log into their own Company A accounts and automate the bulk download of their own files. Company A did not like this. They served me with a highly aggressive, 4-page legal demand (CEASE and DESIST notice). They forced me to:
1. Nuke the automated download tool entirely from GitHub.
2. Remove any mention of their company name from my main open-source project and website (since it’s trademarked). I originally had my tagline as "The Free open-source [Company A\] Alternative," which they claimed was illegally driving their traffic to my site.
3. Remove a feature comparison chart I made. (I admittedly messed up here, I only compared my free tool to their paid tier and omitted their limited free tier, which they claimed was misleading and defamatory).
I'm just a solo dev, so I complied with the core of their demands to stay out of trouble. I scrubbed their name, took down the downloader, and sanitized my website. My main open-source logbook lives independent of them.
I admit I was naive about the legal aspects of comparison marketing and using trademarked names. But the irony is that they probably spent thousands of dollars on lawyer fees to draft a threat against my small project that makes close to zero money (I got a few small donations from happy users).
Has anyone else here ever dealt with corporate lawyers coming after your self-hosted/FOSS projects? It’s a crazy initiation :)
https://redd.it/1rbkx5e
@r_SelfHosted
Self hosted Booking Calendar
Hi everyone. I built a lightweight privacy first self hosted booking calendar for solo operators. It is simple to run, supports localized UI, web push reminders, and ICS export.
Key features
1. Public booking pages and an admin panel for appointments, links and events.
2. Email notifications, web push notifications and webhooks for external integrations.
3. Localized UI with English and Turkish ready.
4. Bun first backend with React and Vite frontend and SQLite via TypeORM for easy hosting.
5. Docker friendly and simple VPS deploy options.
If you want to contribute, please review the project on GitHub, open issues, or submit pull requests, contributions and feedback are welcome.
GitHub: https://github.com/sametcn99/booking-calendar
Docs: https://booking-calendar-docs.vercel.app
https://redd.it/1rb02bt
@r_SelfHosted
when did this sub go downhill
Man, when I first discovered this sub it had a few pieces of spam here and there, but the majority of it was constructive technical discussion and help, and I really do miss those days. These days the most technical anyone seems to get is 'docker exec' and vomiting llm slop and slopware ads. I thought the whole point of self hosting was to get away from the centralization of everything and own your data, and yet every post I see here talking about any kind of software its always just a link to a docker image on a centralized provider and bad build documentation. Maybe this is an old person yelling at clouds moment, but I miss when software required you to know even a little bit about the systems everything runs on instead of being the glorified app store it is now.
https://redd.it/1ray6y0
@r_SelfHosted
PSA: If your self-hosted app uses Cloudflare and you have Spanish users, they might not be able to reach you
Spent hours debugging my production stack thinking something was broken. Turns out all my containers are healthy, TLS cert valid, API responding in milliseconds.
The real problem: Spanish ISPs are blocking Cloudflare IP ranges (188.114.96.x / 188.114.97.x) due to La Liga anti-piracy court orders. Since Cloudflare uses shared anycast IPs, thousands of legitimate sites on those ranges are collateral damage.
Proof:
* From Spain: `ping` [`188.114.97.5`](http://188.114.97.5) → 100% packet loss
* From US: `curl` [`https://mysite.com/health`](https://mysite.com/health) → HTTP 200
* `ping` [`google.com`](http://google.com) from same Spanish network → 0% loss
If you have users in Spain and use Cloudflare, check if your assigned IPs are in the blocked ranges. Worth knowing before you spend hours debugging your stack like I did.
https://redd.it/1ravua8
@r_SelfHosted
A free, open-source self-hosted music composition tool for beat-making, loops, and arrangements - built with Web Audio, WASM, and modern web tech.
It’s a browser-based, Ableton-style music composer designed with a local-first + self-hostable mindset.
What it is
1. A web app for composing music in the browser (clip/track style workflow)
2. Focus: fast UI + tight playback timing (the hard part wasn’t UI, it was scheduling audio reliably)
Why self-hosted users might care
1. You control the hosting + data
2. Works great on a home server behind reverse proxy
3. Can be set up to be LAN-only (perfect for NAS / homelab setups)
Architecture notes (high level)
1. App runs client-side (audio engine uses browser audio primitives)
2. “Server” is mainly for static delivery / app shell (depending on how you deploy)
3. Goal is minimal backend dependencies so it’s easy to host anywhere
Links
Demo: https://composeyogi.com/
Repo + install: https://github.com/AppsYogi-com/ComposeYogi
I’d love your feedback.
https://redd.it/1rapuhi
@r_SelfHosted
MusicGrabber - V2.0.4 released
https://redd.it/1r9s2rn
@r_SelfHosted
Rahoot: A Self-hostable and open-source kahoot
https://redd.it/1ragnql
@r_SelfHosted
SparkyFitness - A Self-Hosted MyFitnessPal alternative now supports PolarFlow & Hevy
https://redd.it/1rac1bw
@r_SelfHosted
Gitea self-hosted for free, includes docker registry too?
I'm just curious, since I'm looking for a self-hosted git server, and I like the idea that some of my projects require building images, and having them automatically be created and available within my homelab, for free would be nice.
https://redd.it/1r6koxo
@r_SelfHosted
Discord Alternatives Comparison
https://redd.it/1r643fj
@r_SelfHosted
How I deployed Matrix
# I've seen countless comments saying how hard it is to deploy Matrix. So here is how I did it.
This setup gives you:
1. Short handles (@me:example.com instead of @me:matrix.example.com)
2. Web Client (Element) included
3. GUI Admin (Synapse Admin) included
4. Federation working out of the box
### 1. The Setup (Don't skip this)
First, create the directories so data persists.
mkdir -p /opt/matrix/synapse
matrix.example.com, your user handles will be long and ugly. We want @user:example.com.docker run -it --rm \
--mount type=bind,src=/opt/matrix/synapse,dst=/data \
-e SYNAPSE_SERVER_NAME=example.com \
-e SYNAPSE_REPORT_STATS=yes \
matrixdotorg/synapse:latest generate
example.com {
# Server-to-Server Discovery
handle_path /.well-known/matrix/server {
header Content-Type application/json
header Access-Control-Allow-Origin "*"
respond `{"m.server": "matrix.example.com:443"}`
}
# Client-to-Server Discovery
handle_path /.well-known/matrix/client {
header Content-Type application/json
header Access-Control-Allow-Origin "*"
respond `{"m.homeserver": {"base_url": "https://matrix.example.com"}}`
}
}
# 2. The Actual Matrix Subdomain
matrix.example.com {
reverse_proxy <matrix:port>
}
services:
matrix:
container_name: matrix
image: matrixdotorg/synapse:latest
restart: unless-stopped
ports:
- "8448:8008"
volumes:
- /opt/matrix/synapse:/data
admin:
image: awesometechnologies/synapse-admin:latest
container_name: synapse-admin
restart: unless-stopped
ports:
- "8449:80"
element:
container_name: element
image: vectorim/element-web:latest
restart: unless-stopped
ports:
- "8450:80"
volumes:
- /opt/matrix/element_config.json:/app/config.json
depends_on:
- matrix
/opt/matrix/element_config.json so the login page defaults to your server:{
"default_server_config": {
"m.homeserver": {
"base_url": "https://matrix.example.com",
"server_name": "example.com"
}
}
}
docker compose up -d
docker exec -it matrix register_new_matrix_user http://localhost:8008 -c /data/homeserver.yaml --admin
example.com.
the fear of the next step
Yeah the title sounds a bit exagerrated, but for myself I am standing in front of a wall and don't know what to do exactly, or how to do. Maybe reddit have some tips. here is my 'problem':
Over the last years I made myself a cool homelab (most of it runs in proxmox) that serves my purposes really well. Besides jellyfin, audiobookshelf and all the media related stuff I also configured nginx as security feature where my services layed behind. My Containers also have proper firewall rulings. I got really confident in hosting services, EXCEPT it incolves 'exposing' my services to some kind of a public. I dont hurt myself on using wireguard as vpn, but this makes me feel kinda unease.
Now, when I think about hosting a minecraft server for my friends, or setting up a Teamspeak-server (related to the bad things Discord did) I hesitate and kinda start overthinking about possible security risks that could come along with it and then stop doing it.
I know about the basic behavior (running non root user, no access from outside for important systems, unattended upgrades etc), and then the next thing I see on this sub is that someone runs a minecraft server on a damn phone.
Do I overthink the whole topic of security? What really is important for hosting a simple dumb private minecraft or ts server for my friends? Everytime this questions come up I feel like hitting a wall of a mass of knowledge I have to acquire before I even dare to try set someting up.
https://redd.it/1r5enr7
@r_SelfHosted