Homelab Atlas: A Self-Hosted Way to Map Out the Homelab My homelab has changed quite a bit over the years, and keeping track of where everything lives isn't always as straightforward as it used to be. A service gets moved to another machine, a new container gets added, and before long, the setup looks nothing like it did six months ago. I usually know where everything is, but trying to explain the whole arrangement to someone else is another story. Homelab Atlas is a self-hosted tool that helps put all of that into a visual diagram. Instead of spending an afternoon drawing boxes and connecting lines, you can add your machines and the services running on them, then let the application do much of the arranging. It's a pretty simple idea, but the finished results look surprisingly good. Putting the Homelab on the Map I've never been particularly excited about making network diagrams. I understand why they're useful, but sitting down with a traditional diagram editor and manually arranging everything isn't exactly how I want to spend my free time. I'd much rather be tinkering with the actual servers. Homelab Atlas makes that process a little more enjoyable. You can create cards for your machines, assign services to them, and have everything arranged into a clean visual layout. It also includes recognizable logos for popular self-hosted applications, which gives the diagrams a much nicer appearance than a bunch of rectangles with text inside them. You can adjust individual cards, add hardware information, and show network connections between devices. That last part could be especially handy for anyone running multiple servers or a collection of mini PCs that have accumulated over time. It's easy enough to remember what's running on a single machine, but once things start spreading across the network, a visual reference becomes a lot more useful. And I'll admit, having a good looking diagram of the homelab is pretty satisfying in its own right. We spend enough time building these things. We might as well have a decent way to show them off. You Don't Have to Enter Everything Manually Homelab Atlas also includes some discovery options that can save you from typing out every service yourself. It can discover Docker containers through an optional collector and identify network devices using Nmap. You can even import results from an existing network scan. I appreciate that discovery doesn't immediately dump everything into your diagram. You review the results and decide what should be included. That's probably for the best, considering how many devices on a typical home network have names that tell you absolutely nothing about what they actually are. It's important to understand that Homelab Atlas isn't trying to be another monitoring dashboard, though. It doesn't constantly watch your containers or automatically update the diagram when something goes offline. This is more about documenting your setup and having a clear picture of how everything fits together. Honestly, I think that's perfectly fine. I've already got enough dashboards telling me what my servers are doing. Sometimes I just want to see where everything belongs. Running Homelab Atlas at Home The developer provides a Docker Compose setup, with container images available for both AMD64 and ARM64 systems. You don't need to configure a separate database, and the basic application works without connecting to any outside service. One detail I would pay attention to is how projects are saved. Homelab Atlas stores your working diagram in the browser, so mounting a Docker volume isn't going to give you the sort of backup you might expect. You'll want to export your project as a JSON file if you plan on keeping it around. You can also export finished diagrams as PNG or SVG images, which makes sharing them pretty straightforward. There's no built-in authentication either, so I'd keep this one on the local network rather than making it publicly accessible. A Good-Looking Homelab Deserves a Good Diagram Homelab Atlas has a really nice appearance for a smaller open-source project. The diagrams are clean without looking overly technical, and the application logos make everything easier to recognize at a glance. I can imagine using an exported diagram in my documentation or pulling one up when explaining how the homelab is arranged. It also doesn't seem interested in becoming some enormous infrastructure management platform, which I find refreshing. It has a fairly specific job, and that's helping you put together a useful picture of your setup without making the process unnecessarily complicated. For anyone whose homelab has grown beyond a single Docker host, I think it's an interesting little project to look into. You can check out the screenshots and Docker installation instructions in the Homelab Atlas GitHub repository. The official documentation has more information about the editor and its discovery features. Homelable: A Visual Map of Your Home Server Setup Homelable brings network discovery, live status checks and diagrams together to help document the machines behind a home server setup. NotedJeremy Explore more on Noted Take a Poll | Play Trivia | Browse Giveaways https://noted.lol/homelab-atlas/ #homelabatlas #way #everything #manuallyhomelab #rss #feeds #linksharing #qwibbit

Narratarr: A Self-Hosted Audiobook Manager With a Built-In Player I've always thought audiobooks deserved a little more attention in the self-hosting world. We've got some excellent applications for organizing movies and television shows, but audiobooks can be a different story. You might have a collection sitting on your NAS, yet finding new books and getting them properly organized often involves more manual work than it should. Narratarr is a new self-hosted audiobook manager that follows the familiar approach of Sonarr and Radarr. It helps you find audiobooks, download them, and organize everything into a library. Unlike the usual *arr applications, though, Narratarr also includes its own audiobook player, so you don't necessarily need another application just to listen to what you've downloaded. Finding Something Worth Listening To I have a habit of browsing through media libraries even when I haven't decided what I want to watch or listen to. Half the fun is discovering something I'd completely forgotten about. Narratarr seems to understand that part of the experience, giving you a proper discovery interface rather than expecting you to arrive with a particular book already in mind. The application connects to LibriVox and Old-Time Radio collections from the Internet Archive, providing access to thousands of freely available recordings. You can browse by genre or search for a particular title, with cover artwork and descriptions helping you figure out what you're looking at. The Old-Time Radio collection is a fun inclusion. Those old radio dramas have a charm of their own, and having them mixed into an audiobook discovery interface makes browsing a little more interesting than another endless list of modern releases. Narratarr also attempts to correct misspelled searches, which is useful when you can't quite remember how an author's name is spelled. That optional feature uses Apple's public search API, so it isn't entirely local, but the application works without enabling it. Getting Books Into the Library Those already running Sonarr or Radarr will recognize the general idea. Find something, add it to the download queue, and let the application handle organizing the files afterward. Downloads from LibriVox and Old-Time Radio are fetched directly over HTTPS, without needing a paid service. Narratarr also supports Prowlarr for searching configured indexers and can pass Usenet downloads to SABnzbd. Torrent downloads currently rely on Real-Debrid, which requires a paid account. That's worth knowing because Narratarr doesn't yet offer direct integration with qBittorrent or Transmission. Thankfully, none of that is necessary to use the free audiobook collections. Finished books are arranged into author and title folders, with cover artwork and descriptive information saved alongside the audio. I can appreciate that because manually cleaning up audiobook downloads isn't exactly an enjoyable afternoon project. Listening Without Another Application Narratarr includes a browser-based audiobook player with chapter navigation and adjustable playback speed. It remembers where you stopped listening, which is especially important when a book takes considerably longer to finish than the average podcast episode. There's a sleep timer as well, including an option to stop at the end of a chapter. That's a thoughtful touch for anyone who enjoys listening before bed and doesn't particularly want to wake up three hours further into the story. The player handles books split across multiple MP3 files alongside M4B audiobooks, treating them as a continuous listening experience. It can also be accessed from a phone through the browser, without requiring a dedicated mobile application. If you're already using Audiobookshelf, there's no need to abandon it. Narratarr can save books into a library Audiobookshelf recognizes and optionally trigger a scan after a download finishes. Running Narratarr at Home The developer provides a Docker Compose configuration with images for AMD64 and ARM64 machines. Everything runs in a single container, with separate mounts for application configuration and the audiobook library. The default web interface uses port 8484. You can optionally enable password protection, although the documentation recommends keeping Narratarr on a trusted home network rather than exposing it directly to the internet. Narratarr is a very new project, so I wouldn't be surprised to see a few rough edges as development continues. The GitHub repository is still relatively small, and there's not much independent feedback from people running it long-term. Still, I think the combination of audiobook discovery and a built-in player makes sense. It's nice having the option to find a public-domain recording and start listening without setting up an entire collection of supporting services first. And for those of us already running an *arr stack, the additional integrations make it easy to see where Narratarr could fit. You can find the Docker Compose setup and screenshots in the Narratarr GitHub repository. The README also explains how the different download sources work, which I'd recommend reading before connecting it to an existing media setup. Explore more on Noted Take a Poll | Play Trivia | Browse Giveaways https://noted.lol/narratarr/ #audiobookmanager #narratarr #player #radio #downloads #something #books #radarr #sonarr #internetarchive #rss #feeds #linksharing #qwibbit

ASRock Industrial AI Box-A395 Review: Taking Strix Halo to Work Today we are taking a look at ASRock Industrial's highly portable Ryzen AI Max+ 395 system, the AI Box-A395. The barebones tower system packs a punch, as well as a handle to quickly pack up the PC The post ASRock Industrial AI Box-A395 Review: Taking Strix Halo to Work appeared first on ServeTheHome. https://www.servethehome.com/asrock-industrial-ai-box-a395-review-taking-strix-halo-to-work/ #asrockindustrial #max #ryzen #ai #rss #feeds #linksharing #qwibbit

PicPeak: Self-Hosted Photo Galleries That Make Sharing Easy Photo libraries and photo galleries aren't quite the same thing. I might want somewhere to keep thousands of family photos, but that doesn't necessarily mean I want to give everyone access to the entire collection. Sometimes I just want to put together a nice album from a particular occasion, send somebody a link, and let them enjoy the pictures without dealing with another account or app. PicPeak is built for exactly that kind of sharing. It's an open-source, self-hosted photo and video gallery platform originally designed with photographers and their clients in mind. But I can see plenty of uses outside professional photography, especially for anyone already running a home server with enough storage to spare. A Better Way to Share an Album Most self-hosted photo applications are focused on organizing an entire collection. PicPeak takes a different approach by treating each gallery as an event. You upload the photos, customize how the gallery looks, and give people a link to view them. You can password-protect galleries and set expiration dates, which is handy when photos are only meant to be shared for a particular period. Visitors can download individual images or grab the entire collection as a ZIP file without creating an account. The interface looks particularly polished, too. Galleries have a clean presentation that feels more like a photography website than a directory full of files. You can customize the branding, and there's support for watermarks if you want to discourage people from casually downloading images. I could see this being useful for sharing photos from a family reunion or putting together an album after a church event. It's also a nice alternative to sending everyone a Google Photos link, especially when the pictures are already sitting on a NAS at home. Let Everyone Add Their Photos One feature that makes PicPeak particularly interesting is guest uploading. Instead of having one person collect everyone's pictures and upload them manually, you can allow guests to contribute directly to the gallery. Anyone who's tried collecting photos after a big gathering knows how that usually goes. Someone sends a few through text messages, another person promises to email theirs later, and the rest are eventually forgotten somewhere on a phone. Giving everyone a place to upload their pictures could make that process a lot less annoying. PicPeak also includes a live slideshow mode that can display incoming photos during an event, which opens up some fun possibilities for gatherings where people are taking pictures throughout the day. For professional photographers, there's another side to all this. Clients can mark favorites and provide feedback, making it easier to decide which photos need further attention. That functionality might be unnecessary for family albums, but it's useful to have when the occasion calls for it. Running PicPeak on a Home Server The developers have made deployment fairly approachable. There's an official all-in-one Docker image that runs PicPeak with SQLite, meaning you don't need to set up a separate database just to experiment with it. For larger installations, there's also a Docker Compose stack using PostgreSQL and Redis. The project supports local storage and S3 compatible backends, so it doesn't require a third-party photo service to handle the actual files. One detail worth mentioning is the ability to reference existing photos from an external media directory without moving the originals. That could be useful for someone who already has a large collection stored on a NAS and doesn't want another application creating duplicate copies. Of course, I'd still make sure the persistent storage is backed up properly. Photo galleries tend to become more important after you've accumulated a few years of them. Not Just for Photographers PicPeak was clearly designed with professional photographers in mind, and some of its more advanced tools reflect that. There's even a customer-management and invoicing system in beta, although those aren't things I'd expect most homelab users to need. What interests me more is the simpler side of the project. It's a good looking way to share a collection of photos without making people jump through hoops. The fact that it runs on a home server makes it even more appealing, particularly when the alternative is uploading a few gigabytes of pictures to another online service. PicPeak is actively maintained and released under the MIT license. The project has also received some encouraging feedback from photographers using it with actual clients, which is reassuring to see beyond the usual GitHub screenshots. If sharing photo albums is something you do regularly, have a look at the PicPeak GitHub repository. There's a live demo available, and the official documentation explains the Docker setup in considerably more detail. Explore more on Noted Take a Poll | Play Trivia | Browse Giveaways https://noted.lol/picpeak/ #rss #feeds #linksharing #qwibbit

Altera Agilex 27 is the First Agilex 2X Platform Adding 800GbE and 1.6TbE and More Altera announced its new Agilex 2X series with the higher-end Agilex 27 with features like LPDDR6 and 1.6TbE support coming in 2027 The post Altera Agilex 27 is the First Agilex 2X Platform Adding 800GbE and 1.6TbE and More appeared first on ServeTheHome. https://www.servethehome.com/altera-agilex-27-800gbe-1-6tbe-hyperflow/ #platformadding #firstagilex #alteraagilex #rss #feeds #linksharing #qwibbit

Fileship: Self-Hosted File Sharing With a Desktop Companion I've been using self-hosted applications long enough to appreciate the little conveniences that make a difference in everyday use. Screenshot uploading is one of them. I don't want to open a file manager every time I need to share an image. I'd rather capture what I need and have it uploaded to my own server without thinking much about it afterward. It's a relatively small part of working with computers, but when you're regularly documenting things or sharing screenshots with other people, those extra steps can become annoying. Fileship is built around that idea. It's an open-source, self-hosted file-sharing application that makes uploading and managing files feel like a natural part of working at a computer. There's a web interface for managing everything, but the desktop integration is what makes the project particularly interesting. Instead of treating uploads as temporary files scattered across different services, Fileship gives them a permanent home on infrastructure you control. A Familiar Idea for ShareX Users I've always appreciated the approach ShareX takes to screenshots. You capture something, it uploads automatically to your configured destination, and you get a link that can be pasted wherever you need it. The catch is that ShareX still needs somewhere to send those files. Public image hosts are convenient, but having an upload destination running on your own server makes a lot more sense when self-hosting is already part of your routine. Fileship approaches that workflow with its own Windows desktop companion, which supports screenshots and screen recording. It also includes OCR for extracting text from captured images, something that can be surprisingly useful when dealing with error messages. Rather than being a direct replacement for ShareX, Fileship combines a familiar capture-and-upload experience with a dedicated server for managing the results. There's also a browser extension, so uploads don't always have to begin with the desktop application. The Windows companion is a nice addition, but Linux users aren't left out entirely. Fileship also provides a Linux shell script for uploading files directly to your self-hosted instance, which could fit nicely into an existing screenshot workflow or a few custom shortcuts. It isn't quite the same polished desktop experience as the Windows application, but for those of us comfortable working with Linux, a simple script can be just as useful. Finding Those Screenshots Again The trouble with screenshot uploading is that it works almost too well. After enough time, you can end up with hundreds of images that were useful for about five minutes before disappearing into a folder somewhere. I've certainly had situations where I remember taking a screenshot of a particular configuration but couldn't remember where I saved it. Filenames consisting of dates and random characters don't help much either. Fileship provides an organized web library where uploads can be browsed and arranged into folders. Shared links can have password protection or expiration settings, which is useful when you only want someone to access a file temporarily. That makes it a little more practical than running a basic upload script that does nothing beyond accepting files and returning URLs. There's also an optional AI-powered search feature that processes images locally on the server. Instead of remembering a filename, you can search based on what appears in the image. For someone who keeps a large collection of screenshots, that could be genuinely useful, although I'd still want a reasonable folder structure rather than depending entirely on AI to find things. Another handy feature is Fileship's Inbox, which lets other people send files directly to your self-hosted instance through an upload link. I can see this being useful when someone needs to send me a collection of screenshots or a large file without going through email attachments. Instead of creating an account for them or dealing with another cloud storage service, I can simply give them somewhere to drop the files and collect everything from my own dashboard. Running Fileship in a Homelab Fileship provides an official Docker Compose deployment using PostgreSQL, so getting it running should feel familiar to anyone already managing a collection of containers. The developer recommends at least two CPU cores and 4 GB of RAM, which is worth considering if you're used to lightweight screenshot uploaders. This is a more involved application, and its resource requirements reflect that. I'd pay particular attention to persistent storage and backups before putting anything important into it. Screenshots might seem disposable when you first upload them, but a collection documenting previous server configurations can become surprisingly valuable. Fileship also supports importing data from Zipline, which may be useful for anyone already running that application and considering a change. Automatic updates are enabled by default, and the optional telemetry setting is another configuration detail worth reviewing during installation. Somewhere Between an Uploader and a File Manager Fileship occupies an interesting spot between a basic screenshot uploader and a larger platform such as Nextcloud. I wouldn't necessarily want to deploy an entire cloud storage suite just to share screenshots, but I also appreciate being able to revisit uploads without hunting through directories on the server. Fileship seems designed around that middle ground, where uploading should be effortless without making the files difficult to manage afterward. Fileship looks surprisingly polished, too. The dashboard is clean and modern, and it doesn't have that unfinished side-project look you sometimes come across when digging through GitHub. It's clear the developer has put some real effort into the interface. Of course, a pretty dashboard doesn't mean everything underneath is perfect, but Fileship certainly makes a good first impression. I'm always happy to see smaller self-hosted projects getting this kind of attention to detail, especially when they're solving a problem many of us deal with regularly. You can find Fileship's Docker Compose setup and more information about the desktop companion in the official GitHub repository. It's worth looking through the screenshots and documentation to see whether the workflow fits how you already capture and share files. Explore more on Noted Take a Poll | Play Trivia | Browse Giveaways https://noted.lol/fileship/ #desktopcompanion #screenshot #file #linux #sharex #idea #familiar #rss #feeds #linksharing #qwibbit

Noctua Fan Swap on a Calnex Rb/GPS Frequency Reference Since this will be the only source of this process on the Internet1, I figured I'd document a quick fan swap I did on a piece of test/reference equipment I'm running in my Atomic Clock Tower. I took my Torx T10 screwdriver and popped off the top cover. The screws required some force to loosen, as they were all secured with Loctite. The inside is simple and serviceable. There's a SpectraTime GRClok-1500, with an integrated timing-grade GPS module, and a stabilized 10 MHz output2. https://www.jeffgeerling.com/blog/2026/noctua-fan-swap-calnex-rb-gps-frequency-reference/ #fanswap #gps #reference #clocktower #atomicclock #torx #loctite #mhz #rss #feeds #linksharing #qwibbit

Microsoft Maia 300 Shown at OCP 2026 At OCP Summit 2026, we saw the new Microsoft Maia 300 platform with four 288GB HBM4 950W accelerators per tray The post Microsoft Maia 300 Shown at OCP 2026 appeared first on ServeTheHome. https://www.servethehome.com/microsoft-maia-300-shown-at-ocp-2026/ #microsoftmaia #ocpsummit #rss #feeds #linksharing #qwibbit

Hold My Docs: A Self-Hosted Wiki That Keeps Everything in Git I've lost count of how many times I've made changes to a server and told myself I'd document everything later. A few months go by, another container gets added, something breaks, and suddenly I'm trying to remember why I configured a service a particular way. Keeping good documentation is one of those habits that pays off eventually, even if sitting down to write it isn't always the most exciting part of maintaining a homelab. Hold My Docs is a self-hosted wiki that takes an interesting approach to keeping those notes organized. Instead of hiding everything inside a database, it stores documentation as ordinary Markdown files in a Git repository. You get a proper web interface for writing and browsing, but the underlying documents remain accessible without the application. A Wiki With a Memory One of the more useful things about Hold My Docs is how it handles revisions. Every time you save a page, the application creates a Git commit. That gives each document a history you can revisit when something changes or an older set of instructions turns out to be the one you actually needed. For homelab documentation, I can see this being especially handy. Network configurations change, Docker stacks get reorganized, and sometimes a perfectly good setup gets replaced by something that seemed like a better idea at the time. Being able to look back at previous instructions is a lot nicer than digging through old backups. The application can also synchronize with a remote Git repository, making it possible to maintain another copy of your documentation outside the wiki itself. More Than a Collection of Markdown Files Hold My Docs includes a browser-based Markdown editor with previews, internal links, and backlinks between pages. It also supports Mermaid diagrams, which could be useful for mapping out how services communicate or documenting a network without needing separate diagramming software. Documentation can be divided into namespaces, each with its own navigation and visibility settings. That makes it possible to keep server administration notes separate from household information or other material you might want to share. There's full-text search for finding information across wiki pages. An optional Apache Tika integration extends searching to supported PDF and Office documents, although that requires an additional service. I think the real benefit here is having everything in one place without making the application itself the only way to access your work. Markdown files are easy to move around, and Git provides a familiar way to manage their history. Running Hold My Docs on Your Server Hold My Docs is written in Go and can run as a standalone binary, but the developer also provides an official Docker Compose configuration. The example deployment listens on localhost by default, which is a sensible starting point for a service that might contain private infrastructure details. Authentication supports local accounts and OIDC, making it possible to integrate with an existing identity provider. Administrators can manage access through scoped permissions and personal access tokens. The documentation also explains how to configure the initial administrator account and remove the bootstrap credentials afterward. One storage detail is worth paying attention to. The application's data directory should remain on local storage rather than NFS, although the Git repository itself can live on network storage. That's useful to know before deciding where to mount persistent volumes. Your Documentation Doesn't Have to Stay in the Wiki I really like how you can export a namespace as a static website. You could maintain documentation privately and publish selected sections without requiring visitors to access the wiki application itself. There's also an optional MCP server for allowing authorized AI tools to interact with documentation. That could have practical uses for searching through a large collection of notes, although I'd be careful about granting access to anything containing private infrastructure information. Hold My Docs is released under the MIT license, and its developer recommends maintaining backups and testing upgrades before depending on it for important documentation. What I like is that the application doesn't try to reinvent how documents are stored. It puts a convenient interface around Markdown and Git, both of which are perfectly capable of surviving long after a particular web application disappears. I've always thought the best homelab documentation is the kind you can still read when the server hosting your documentation is the very thing you're trying to fix. Hold My Docs seems to understand that distinction. You can find the source code, screenshots, and Docker installation instructions in the Hold My Docs GitHub repository. The official documentation covers authentication, storage configuration, and the other details worth reviewing before adding it to your server. Explore more on Noted Take a Poll | Play Trivia | Browse Giveaways https://noted.lol/hold-my-docs/ #docs #hold #git #markdownfiles #documentation #serverhold #docker #runninghold #rss #feeds #linksharing #qwibbit

ASRock Creator Radeon AI Pro R9700 Review: Workstation Inference, AMD Style Today we are taking a look at ASRock's high-end Radeon AI inference card for workstations, the Creator Radeon AI Pro R9700. With 32GB of RAM and a $1700 price tag, the card is a highly affordable way to fill out a workstation for local AI inference The post ASRock Creator Radeon AI Pro R9700 Review: Workstation Inference, AMD Style appeared first on ServeTheHome. https://www.servethehome.com/asrock-creator-radeon-ai-pro-r9700-review-workstation-inference-amd-style/ #creatorradeon #pro #amd #inference #ai #rss #feeds #linksharing #qwibbit