Magnetissimo is a self-hosted torrent indexer that searches eight public sources at once and remembers everything they return, so the release you hunted for across half a dozen sites last week is already waiting in your own index when you look for it again. A search answers from the local database immediately and asks YTS, Nyaa, Sukebei, TorrentDownloads, AudioBookBay, The Pirate Bay, EZTV, and OxTorrent in parallel; every answer is filed into the index as it lands and streams into the results page with size, seeders, leechers, and age attached. The query language goes further than any single source: exact quoted text, exclusions, alternatives, wildcards, and filters for file size, seeder count, and age, with matching that copes with dotted scene release names and titles without spaces. An optional DHT crawler walks the peer-to-peer network in the background at a rate you choose, discovering torrents and their file lists around the clock and showing arrivals live on its own page. Media automation connects natively: the Torznab endpoint feeds Sonarr, Radarr, Lidarr, Prowlarr, and book managers with standard categories, and a JSON API serves the latest and best-matching torrents to scripts. Everything lives in one SQLite database inside a single container, search and paging work without JavaScript, and upgrades are a pull and a restart. Running on a dedicated VPS on RepoCloud with guaranteed CPU, RAM, and SSD, full root SSH access, and a browser serial console. MIT licensed.

Hocuspocus is the WebSocket backend that turns a regular editor into a multiplayer one: two people open the same document, type at the same time, and watch each other's words appear in place, with every keystroke merged cleanly and no save conflicts to untangle later. The server speaks the Yjs synchronization protocol, so clients built with Tiptap, ProseMirror, Slate, Quill, or Monaco editors connect over a persistent connection and exchange small updates that the conflict-free data type orders correctly regardless of arrival sequence, and awareness messages carry cursor positions and presence so collaborators see who is in the document. Teams that outgrow one process add the Redis extension to balance connections across multiple server instances, reaching very large numbers of simultaneous editors on the same documents. Persistence extensions store documents in SQLite for small deployments or in a database or S3-compatible storage through your own fetch and store callbacks, the webhook extension notifies your application when documents change, throttling guards against abusive connections, and hooks for authentication and document loading let you enforce per-document permissions before anyone joins. Developers embed the server into an existing Node.js, Bun, Deno, or Cloudflare Workers runtime, or run it as the collaboration tier beside their main application, and offline-first clients queue changes locally and synchronize when the connection returns. Running on a dedicated VPS on RepoCloud with guaranteed CPU, RAM, and SSD, full root SSH access, and a browser serial console. MIT licensed.

Bluesky PDS is the official Personal Data Server for hosting your own account on the AT Protocol network, so your posts, follows, and media live on infrastructure you control and still appear across the whole federated Bluesky social graph. An interactive installer sets up the container stack with the Caddy web server handling HTTPS for your domain and wildcard subdomains, creating the data directory, generating secrets, and registering the service with the network's identity directory. You create accounts directly or by invite code, sign in from the standard Bluesky web, iPhone, and Android apps by pointing them at your server address, and brand the sign-in, sign-up, and account screens with your own service name, logo, colors, and background images. Email verification and password resets go through your SMTP provider, rate limiting and blob upload limits are configurable, and the data directory holds everything needed to back up or migrate the instance to a new host later. Built-in tooling covers day-to-day operation: an admin command updates the service, automatic updates run through Watchtower, logs stream to Docker or a file, health and firehose WebSocket endpoints confirm the network can reach you, and an optional Prometheus and Grafana stack tracks accounts, sign-ins, and request latency. Account migration moves an existing identity onto your server when you are ready to leave a hosted provider. Running on a dedicated VPS on RepoCloud with guaranteed CPU, RAM, and SSD, full root SSH access, and a browser serial console. MIT and Apache-2.0 licensed.

The invoices, reconciliations, and report downloads your team repeats every week are exactly the work a robot should be doing at 2 a.m. Astron RPA is an open-source robotic process automation platform: business users assemble workflows in a visual designer from more than three hundred pre-built components for clicking through desktop applications, driving browsers, reading and writing Excel and PDF files, querying databases, and calling APIs, then robots execute those workflows on schedule, on demand, through API calls, or through MCP services. The platform covers Windows desktop software and web pages alike, reaching office suites, browsers, and the finance and ERP systems that never grew an API, and Python scripting extends any step the components do not cover. A built-in excellence center and team marketplace let organizations publish shared automations, manage permissions, monitor executions, and reuse each other's robots across departments instead of rebuilding the same process ten times. AI components and a native agent platform connect language models to the workflows, so an agent can reason about a task and hand the clicking to a robot, and a workflow can call the agent back when judgment is needed. Self-hosters run the control plane themselves with Docker: database, identity service, and web console behind their own domain and certificates. Running on a dedicated VPS on RepoCloud with guaranteed CPU, RAM, and SSD, full root SSH access, and a browser serial console. Apache-2.0 licensed.

Your monitoring should be able to tap you on the wrist, not just write to a log nobody reads. Bark is an open-source push notification server for iPhone and iPad: any script, cron job, or monitoring tool sends one HTTP request to your server, and a custom notification lands on your iOS device through Apple's push service, usually within a second and without draining the battery. Notifications carry titles, subtitles, bodies, badges, custom icons, and sounds, and the iOS-only vocabulary is fully supported: grouping related alerts together, time-sensitive delivery that breaks through Focus, critical alerts that keep ringing for thirty seconds, and tap actions that open a URL or copy a value to the clipboard. Batch keys notify several devices at once, archived messages can carry a time-to-live before they delete themselves, and encrypted pushes keep content unreadable to anyone in the middle. The server itself is deliberately tiny: one container or binary listening on a port, device tokens stored in a local file database by default, an endpoint to register devices, health checks for your supervisor, and even an MCP interface so AI agents can notify you through the same pipe. Cron jobs, backup scripts, CI pipelines, and home automations all gain a voice that reaches the phone in your pocket. Running on a dedicated VPS on RepoCloud with guaranteed CPU, RAM, and SSD, full root SSH access, and a browser serial console. MIT licensed.

Spreadsheets deserve an engine that anyone can embed, audit, and run on their own server. IronCalc is an open-source spreadsheet engine with a browser editor that reads and writes real Excel files: open an xlsx, edit values and formulas in the grid, format cells, and save the workbook back without a round trip through a proprietary suite. The grid feels familiar on purpose, with a formatting toolbar for fonts, fills, borders, number formats, and alignment, sheet tabs along the bottom, and templates for invoices, expense trackers, calendars, timesheets, and project plans so a new workbook starts from structure instead of a blank page. Files import from xlsx and export again to the same format, which keeps IronCalc honest as a working tool instead of a viewer. Underneath, the calculation engine is a library other software can call, so the same formulas that evaluate in your browser can also run inside a Python script, a Node service, or a WebAssembly build embedded in your own product. Self-hosting puts the whole webapp behind your own address: one Docker Compose brings up the spreadsheet server with its proxy, and teams share workbooks from infrastructure they control rather than uploading budgets to a stranger's cloud. Running on a dedicated VPS on RepoCloud with guaranteed CPU, RAM, and SSD, full root SSH access, and a browser serial console. MIT and Apache-2.0 dual licensed.

Your own hardware should be the cloud your AI lives in. Olares is an open-source personal cloud operating system that turns one machine, or a small cluster of them, into a private AI platform: run local large language models, keep always-on agents working on your files, and open a browser-based desktop of apps from any device. The built-in Lares assistant takes instructions in plain language and delegates research, file management, and routine automation to agents. The Model Console downloads and serves open models from a local catalog, with GPUs pooled across nodes and split by time or memory for AI, media, and gaming workloads. A one-click market installs self-hosted applications beside the system apps: Files browsing local folders, synced data, cloud storage, and SMB or NFS shares; Vault managing secrets; Dashboard watching the machine; and Control Hub administering it all. Networking stays private by design, with a built-in VPN, reverse proxy, and public, private, or internal entrances giving each app an HTTPS address without exposing individual ports. Backups protect the whole system, the desktop works from phone or browser anywhere, and your apps, storage, and accelerators behave like one coherent computer rather than a pile of containers. Running on a dedicated VPS on RepoCloud with guaranteed CPU, RAM, and SSD, full root SSH access, and a browser serial console. AGPL-3.0 licensed.

Documentation deserves better than a database nobody wants to maintain. Typemill is an open-source flat-file CMS for informational websites, manuals, knowledge bases, and handbooks: content lives as Markdown files in folders, so a site can be copied, versioned, backed up, and moved with ordinary file tools and no database server anywhere in the stack. Authors get a visual block editor for nontechnical writing and a raw Markdown editor when precision matters, with a navigation tree, drag-and-drop structure, media library, page metadata, and publish states in one workspace. The same source then publishes three ways: as a themed website, as a printable PDF manual, or as an EPUB eBook through the eBook studio, making single-source publishing real for teams tired of maintaining docs, PDFs, and handbooks as three diverging copies. Themes are built from HTML and Twig; plugins add search, analytics, diagrams, math, and integrations; and an optional AI assistant interface connects writing help to the editor. User management, restricted pages, redirects, and a REST API cover the operational side, and the whole application is small enough to run comfortably on modest shared-style hosting or a tiny server while still feeling instant. Running on a dedicated VPS on RepoCloud with guaranteed CPU, RAM, and SSD, full root SSH access, and a browser serial console. MIT licensed.

Testing a migration against a toy database proves almost nothing. Database Lab Engine is an open-source platform that clones production-sized PostgreSQL databases in seconds, giving every developer, CI job, and query experiment a full-fidelity copy without waiting hours for a restore or paying to store ten of them. The trick is thin cloning: copy-on-write snapshots taken on ZFS or LVM let a one-terabyte database materialise as a writable branch in about ten seconds, with dozens of independent clones sharing one physical dataset until their changes diverge. Teams point CI at a fresh clone for every migration, so 100 percent of schema changes get rehearsed against real data shapes and volumes before they touch production; analysts get disposable copies for testing query rewrites and index ideas; and a bad experiment costs one delete, not one restore. Clones are managed through a web UI, API, and CLI, draw from self-managed Postgres or managed sources on RDS, Cloud SQL, and Supabase, and carry popular extensions in their containers. Resets, upgrades, deletion protection, retention rules, and resource quotas keep the fleet tidy as it grows into branches and snapshots numbering in the dozens. Running on a dedicated VPS on RepoCloud with guaranteed CPU, RAM, and SSD, full root SSH access, and a browser serial console. Apache-2.0 licensed.

One article should not require a subscription, an account, and your tracking consent. 13ft is a self-hosted alternative to the defunct 12ft.io that fetches web articles the way a search-engine crawler receives them, then shows you the clean page without paywall walls, ad clutter, and pop-up nonsense attached. Paste a link into the single input, or better, install the companion bookmarklet: one click on any article page rewrites the address to your own server, and the reading copy comes back from infrastructure you control. Because it is yours, there is no third-party service logging every article you read, no central site that can be taken down and strand the habit, and nobody selling a premium tier of unblocking. It handles the stubborn cases with a loading page and multiple fallback retrieval methods that kick in when the first approach fails, ships localised interface strings in several languages, and the whole thing is one small application that runs from a container and sits politely behind a reverse proxy. Casual readers reclaim the occasional long-form piece; researchers and archivists get a reproducible way to fetch the same copy a search index saw. Running on a dedicated VPS on RepoCloud with guaranteed CPU, RAM, and SSD, full root SSH access, and a browser serial console. MIT licensed.

Your IPTV provider's login should never be typed into six televisions. IPTV Proxy is an open-source reverse proxy that takes an M3U/M3U8 playlist or an Xtream Codes account and re-serves every channel, movie, series, and guide from your own address, under a username and password you choose. Players like TiviMate, IPTV Smarters, VLC, and Kodi talk only to the proxy, so the provider's real URL and credentials stay hidden on one machine, and when the provider changes its address, your household devices keep pointing at the same stable endpoint behind your domain, HTTPS, or VPN. Each viewer can receive their own proxy account with a connection limit, regex filters trim channel lists down to what a household actually watches, playlist refreshes happen automatically, and a read-only status page shows who is watching what right now. Media servers join in too: the proxy presents live channels as an HDHomeRun network tuner, so Plex, Jellyfin, Emby, and Channels DVR record and time-shift them with guide data included. There is no database and no admin panel to secure; one small binary or container, configured by flags, is the whole deployment. Running on a dedicated VPS on RepoCloud with guaranteed CPU, RAM, and SSD, full root SSH access, and a browser serial console. GPL-3.0 licensed.

Some data refuses to sit still in rows and columns. Skytable is an open-source NoSQL database that stores information in spaces and models instead of rigid tables, letting a record carry strings, numbers, nested lists, and other complex types the way your application actually thinks about them. Developers query it with BlueQL, a language that feels familiar to anyone who writes SQL but adds modern collection syntax and mandatory parameterisation that blunts injection attacks before they start. Underneath, the engine keeps working data in memory, drives multithreaded asynchronous network I/O, batches writes, and persists through a custom append-only storage layer with delayed-durability transactions that trade a tunable sliver of risk for a lot of speed. Client drivers connect applications directly, an interactive shell handles administration and exploration, authentication protects the root account and additional users, and clustering work in progress aims at the high-availability deployments larger systems expect. Teams reach for it when a document store feels too shapeless, a relational database feels too stiff, and latency budgets leave no room for either. Running on a dedicated VPS on RepoCloud with guaranteed CPU, RAM, and SSD, full root SSH access, and a browser serial console. AGPL-3.0 licensed.

Security teams drown in alerts that all demand the same twelve manual steps. Shuffle is an open-source security orchestration, automation, and response platform that turns those steps into visual workflows: an alert arrives from the SIEM, Shuffle enriches it against threat intelligence, opens the ticket, isolates the host, and messages the analyst, without anyone copying values between browser tabs. Analysts assemble workflows on a canvas from more than two thousand pre-built app integrations covering email, EDR, firewalls, ticketing, chat, and cloud services, and any tool with an OpenAPI description can be converted into a new app automatically, exported again, and reused outside Shuffle. Webhooks, schedules, and pipeline triggers start executions; custom Python covers the odd cases; conditional branches and sub-workflows keep one big playbook from becoming spaghetti. Multi-tenancy with sub-organisations, roles, single sign-on, and MFA fits managed security providers running one deployment for many customers, and execution workers can run on premises closer to the systems they touch. Community members publish workflows and apps to a marketplace, so a phishing triage playbook that works for one SOC can be running in another by lunchtime. Running on a dedicated VPS on RepoCloud with guaranteed CPU, RAM, and SSD, full root SSH access, and a browser serial console. AGPL-3.0 licensed.

The front end should never dictate how your content is shaped. Squidex is an open-source headless CMS that stores content once and delivers it to any website, app, or service through REST, OData, and GraphQL APIs, so teams design schemas around their domain instead of around a theme. Editors work in a clean management UI: define schemas with validation and localisation, write content with references and components, manage images and files with resizing built in, and move drafts through custom workflows that say who creates, reviews, and publishes. Nothing waits on a developer for routine changes, and nothing ships unreviewed in organisations that need approval chains. Rules react to every content event, pushing fresh posts to syndication partners, notifying search indexes, invalidating caches, or calling your own webhooks; backups, versioning with diff-style history, and role-based permissions keep an auditable operation. Multiple projects live in one installation, clients authenticate API consumers, and content plugs into static-site generators, storefronts, and mobile apps from one hub. Because you run it yourself, the database options span MongoDB and several SQL servers, and the API your applications depend on answers from infrastructure you control with no per-seat tax on editors. Running on a dedicated VPS on RepoCloud with guaranteed CPU, RAM, and SSD, full root SSH access, and a browser serial console. MIT licensed.

Five monitoring systems should not mean five places to look when something breaks at 3 a.m. Alerta is an open-source alert management server that consolidates alerts from many sources into a single stream, de-duplicates repeats, correlates severity changes, and shows only the current state of each problem in a web console built for fast scanning. Prometheus, Zabbix, Nagios, InfluxDB Kapacitor, cloud watch alarms, and your own scripts all post into the same HTTP API, which accepts a flexible alert format carrying environment, resource, event, severity, and custom attributes, so no tool has to contort itself into someone else's schema first. Repeats of a known fault raise a duplicate count instead of generating another page, and a severity change updates the existing alert rather than birthing a twin. Operators acknowledge problems, assign them, silence expected fire during blackout periods set for maintenance windows, and partition visibility by customer for teams that monitor other people's infrastructure. A command-line client, query language, webhooks, and plugin hooks carry alerts onward into tickets, chat, and paging workflows, and everything persists in MongoDB or PostgreSQL with API keys and OAuth protecting access. Running on a dedicated VPS on RepoCloud with guaranteed CPU, RAM, and SSD, full root SSH access, and a browser serial console. Apache-2.0 licensed.

Dashboards stop being screenshots of someone's browser and start being code your team reviews. Perses is an open-source dashboard platform for observability data, displaying Prometheus metrics, Tempo traces, Loki logs, and Pyroscope profiles in one interface, with more datasources arriving through plugins. Teams build dashboards in the visual editor, organising panels into groups with variables that retarget every chart at a new job, instance, or environment; the same dashboards can be written with Go or Cue SDKs, validated in CI, and applied through the CLI or API, so a hundred services keep a hundred consistent dashboards under version control. Projects organise ownership, role-based access controls who can edit what, and authentication connects to the identity provider you already run. Because the dashboard model is an open specification, definitions stay portable instead of locked in one vendor's JSON dialect, and the panel libraries can even be embedded inside your own product's interface. A Kubernetes operator mode stores dashboards as custom resources inside application namespaces, so the team that owns a service owns its dashboard next to its deployment. Time ranges, refresh controls, and explore views make it as useful during an incident as it is as long-lived documentation of how a system behaves. Running on a dedicated VPS on RepoCloud with guaranteed CPU, RAM, and SSD, full root SSH access, and a browser serial console. Apache-2.0 licensed.

A social network shaped like your community, not like an advertiser's funnel. Bonfire is an open-source toolkit for building federated online spaces where a group hosts its own home, sets its own rules, and still talks with the wider fediverse. Pick a flavour to start: Social for familiar timelines and profiles, Community for groups and topics, Open Science for research collaboration, with coordination and cooperation spaces in development. Everything is an extension, so each instance enables only the pieces it wants, from boundaries and circles that control exactly who sees each post, to custom feeds, direct messages, bookmarks, and trending topics. Members follow people and groups across servers, boost and quote posts, and carry one identity through conversations that cross instance lines. Moderation is granular by design: instance, group, and personal controls combine, and public posts can stay readable without an account to welcome newcomers. Organisers customise branding, extensions, and defaults instead of accepting a platform's roadmap, and developers write new extensions when a community needs a feature nobody built yet. The result is a small, governable internet: your server, your members, connected to everyone. Running on a dedicated VPS on RepoCloud with guaranteed CPU, RAM, and SSD, full root SSH access, and a browser serial console. AGPL-3.0 licensed.

Spotify gets a Wrapped; your media server deserves one too. Rewindarr is an open-source companion for Plex and Jellyfin that turns play history into a story your users step through, card by card: total watch time, how it split across movies, shows, and music, how they rank against the average viewer, and the viewing personality their habits earn them, ending in a summary they can save as an image and share. The same deployment doubles as an everyday dashboard for the whole server, with watch time, requests, library size, and per-library leaderboards for movies, shows, music, and users across a week, a month, a year, or all time. Live activity shows what is streaming right now, and searchable history groups binges together so a weekend disappears into one tidy entry. Ratings from TMDB, IMDb, and Rotten Tomatoes sit beside every title, request buttons reconnect viewers with deleted favourites through Seerr, and anonymisation trims user details when you share the dashboard wider. Statistics flow in from Tautulli or Tracearr for Plex and from Tracearr for Jellyfin, everyone signs in with the Plex or Jellyfin account they already have, and the installable app works on the TV-side phone as happily as on a desktop. Running on a dedicated VPS on RepoCloud with guaranteed CPU, RAM, and SSD, full root SSH access, and a browser serial console. GPL-3.0 licensed.