Kokoro-FastAPI turns text into natural-sounding speech across eight languages by serving the 82-million-parameter Kokoro-82M model through an OpenAI-compatible REST API, so any existing OpenAI SDK client can generate audio by just changing the base URL. With over 5,300 GitHub stars since December 2024, the fully Dockerized FastAPI server covers American English, British English, Spanish, French, Hindi, Italian, Japanese, Brazilian Portuguese, and Mandarin Chinese with language-specific phoneme processing. Inline voice mixing blends multiple profiles using weighted ratios like af_bella(2)+af_heart(1), automatically normalizing weights and caching combined voicepacks as PyTorch tensor files for reuse. Audio streams in real time over HTTP with configurable chunk sizes, or generates complete files in MP3, WAV, OPUS, FLAC, AAC, or PCM formats with speed control from 0.25x to 4.0x. Per-word timestamped captions with speaker-tagged voice labels enable subtitle generation for podcasts, audiobooks, and accessibility workflows. Pre-built Docker images support NVIDIA GPU acceleration via CUDA, experimental AMD GPU inference via ROCm, and CPU-only deployment on linux/amd64 and linux/arm64 architectures, with Apple Silicon MPS support available through direct UV execution. The integrated web interface at port 8880 provides browser-based speech generation, while the Swagger UI at /docs exposes the full API reference. Debug endpoints report system statistics for monitoring inference load. 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 library for the books you read and the books you listen to: Chaptarr manages audiobooks and eBooks side by side, so the same title can live on your server as a chaptered M4B and an EPUB without two separate apps fighting over it. It knows the difference between editions and narrators, which means a dramatized production, an unabridged reading, and the print text can coexist as first-class versions of one work. Multi-part releases and multi-file recordings are handled as single books, and MP3 audiobooks can be converted into one chaptered M4B with chapters preserved or inserted where they are missing. Separate root folders keep audiobooks and eBooks apart, or place eBooks alongside their audio twins; either way, matching uses file tags, folder names, and release metadata to attach files to the right book automatically. Series are organized for you, metadata profiles control which languages belong in your library, and flexible renaming tokens produce filenames that read the same way a human would shelve them. Custom quality profiles per library drive automatic upgrades when a better release appears, and the standard download clients and indexer integrations from the *arr family plug straight in. 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.

LMS is a self-hosted music server that streams your own library to any web browser or Subsonic-compatible player. Point it at a folder of tagged files and it builds a library that understands multi-valued genres and moods, composer, conductor, and producer credits, release types, and release groups that keep remasters beside the original. MusicBrainz identifiers settle duplicate artist and album names, and an artist information folder in Kodi NFO format adds biographies and portraits. Browse with filters like Electronic plus Relaxed, then switch on radio mode and the queue refills itself with similar tracks, chosen by audio similarity or by tags. Plays can be scrobbled to ListenBrainz and Last.fm, and loved tracks synchronize with ListenBrainz. Phone apps connect through the Subsonic and OpenSubsonic API with a personal API key, and transcoding trims bandwidth on slow connections. ReplayGain evens out loudness, podcasts and lyrics sit in the same interface, and an admin-only jukebox mode plays through the server's own audio output. Several libraries can live side by side, a .lmsignore file keeps unwanted folders out of scans, and logins can use the internal user database, PAM, or reverse-proxy SSO headers. 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.

A folder of loose audiobook chapters goes in; a single, perfectly tagged m4b library comes out. Bragibooks is the operations board for that transformation: you select a batch of audiobook folders from your server, and it searches AudiobookDB to match each one with the right title, author, narrator, and release, with a custom search on hand when the first match is wrong. Confirmed books enter a processing queue where m4b-merge merges split mp3, m4a, and m4b parts, converts formats, strips stale metadata, and writes clean, consistently tagged files while you watch the live job output in your browser. Saved output profiles let you route different batches to different directories or path formats, and the books library keeps every finished title with its metadata, files, and processing history a click away. The whole thing runs as one container with the interface built in, a SQLite database on a mounted config volume, and access restricted to your local network by default, with an optional API key for external tools. Audiobook collectors use it to tame download folders before importing into their player of choice, and hoarders of split-chapter rips finally get uniform Chapter 01 style files. 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.

Minimax Music UI turns a self-hosted MiniMax-Music3 model into a private song factory, giving you a Suno-style studio where you describe a track, add lyrics, and watch finished songs appear in your own library. The Create panel works in Simple or Custom mode: write a style description covering genre, BPM, key, vocals, and arrangement, then shape words in a lyrics editor with one-tap [Verse], [Chorus], [Bridge], and [Instrumental] section chips. One submit can queue up to four takes at 5 to 360 seconds each, and the render queue shows queued, rendering, and done states with elapsed time, cancel, and retry, surviving a server restart without losing your place. When you run the bundled inference server, you can even press play on a track that is still rendering and hear the waveform grow in real time. A magic wand rewrites rough ideas into structured music captions, and a second wand drafts original lyrics sized to your chosen duration, both through any OpenAI-compatible model you point it at. Finished renders land in a searchable library with cover art, seeds, and render times, ready to download as WAV, reuse as a template, or open in a detail panel with the prompt and lyrics. A sticky waveform player with seek and keyboard shortcuts keeps music going while you work, and an optional access token locks the whole studio down. Running on a dedicated VPS on RepoCloud with guaranteed CPU, RAM, and SSD, full root SSH access, and a browser serial console. MIT licensed.

BookLore centralizes personal electronic book, audiobook, and comic collections into an organized private media server equipped with automated metadata scrapers and e-reader synchronization. Users can drop EPUB, PDF, CBZ, and MOBI files into a watched BookDrop directory to trigger automated format parsing, background cover retrieval, and staged queue imports. Integrated metadata scrapers query Google Books, Open Library, Goodreads, and Amazon to automatically fill synopsis summaries, author records, publishing dates, and review scores. Dynamic Magic Shelves categorize volumes through custom rule-based filters, author collections, and full-text search indexes across your complete literature catalog. The browser-based reader renders documents with adjustable typography, night modes, audio playback controls, text annotations, and persistent bookmarking. Hardware e-readers and mobile devices connect through native Kobo Store emulation APIs, bidirectional KOReader progress synchronization, and OPDS catalog feeds. Multi-tenant permissions provide each household member with private reading statistics, personalized shelves, email book delivery, and direct send-to-Kindle dispatch. Running on a dedicated VPS on RepoCloud with guaranteed CPU, RAM, and SSD, full root SSH access, and a browser serial console. GNU AGPLv3 licensed.

AzuraCast is a solution for running internet radio stations, packaging the entire broadcast stack into a single Docker installation that gets you on air in minutes. Built on PHP with a Vue.js frontend, it uses Liquidsoap as the AutoDJ engine to compile media from playlists, live DJ inputs, and scheduled programming into composited output streams, then broadcasts via Icecast-KH, SHOUTcast 2 DNAS, or Rocket Streaming Audio Server to thousands of concurrent listeners. The media library supports drag-and-drop uploads with automatic metadata extraction, while playlist management handles sequential, random, weighted, and time-scheduled rotations. Live DJ accounts let remote broadcasters connect via BUTT, Mixxx, or any Icecast-compatible software, with automatic crossfade transitions between AutoDJ and live input. The built-in Web DJ tool enables broadcasting directly from the browser without additional software. Each station includes public-facing player pages with now-playing metadata, album art, song history, and embeddable widgets for external websites. The listener request system lets audiences queue songs through the public interface. Multi-station administration hosts unlimited stations on a single server with individual user accounts and granular role-based permissions. Comprehensive analytics track listener counts, geographic distribution, unique listeners, and listening duration with exportable reports. HLS streaming delivers adaptive bitrate broadcasts for mobile compatibility. LADSPA audio plugins enable professional equalization and processing in the stream pipeline. 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.

Narratarr finds free audiobooks, fetches them through your download stack, and plays them back in the browser, so a book goes from curiosity to chapter one in a single click. Discovery starts with more than 30,000 legal titles from LibriVox and Old-Time Radio, browsable by genre, popularity, language, and format. Search forgives your spelling: type a mangled author name and the right books still surface, ranked by how well they match. Every title shows its runtime, format, size, and description before you commit, and one click queues the download, files it under Author and Title with cover art attached, and tracks progress live on the Activity page. Playback needs no second app. The built-in player handles chapters, including markers buried inside m4b files, plays from three-quarter speed to triple speed, and offers a sleep timer that can stop at the chapter boundary. Your place is kept per listener, so a household can share one library without losing anyone's bookmark. Follow authors and series on the Wanted list and new releases arrive on their own; notifications land in ntfy, Gotify, Discord, or any webhook when downloads start and finish. Ebooks ride along too, readable as EPUBs in the same interface. Running on a dedicated VPS on RepoCloud with guaranteed CPU, RAM, and SSD, full root SSH access, and a browser serial console. MIT licensed.

Drop a hard drive's worth of comics, e-books, and audiobooks into folders, point Kumiho at them, and the whole collection becomes a library you can browse and read from any browser. The scanner walks nested folders recursively and treats the folders that actually hold pages or files as series, so an existing collection works as it is, with no metadata matching ritual before the first read. Comics and manga open in a responsive viewer built for the format: page turning on a desktop, continuous webtoon scrolling on a phone or tablet. E-books in EPUB, PDF, and TXT open in the same library, while archives in ZIP and CBZ are read directly without unpacking. Audiobooks get a real player, with a chapter list, resume playback that remembers where you stopped, progress tracking, bookmarks, and a sleep timer for listening at night. A series folder can even carry its own soundtrack: an MP3 whose filename matches starts playing as background music while you read. Everything runs from one lightweight Go server in a single Docker container, storing its database and configuration in mounted volumes, so the library sits comfortably on a small VPS next to the files it serves. 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.

With 26,000 GitHub stars and consistent victories over ElevenLabs in blind evaluations, Chatterbox delivers state-of-the-art text-to-speech with zero-shot voice cloning requiring only 5 seconds of reference audio. The model family spans three architectures: Chatterbox Multilingual V3 (500M parameters, 23+ languages including Arabic, Chinese, Japanese, Korean, Hindi, French, German, Spanish, and Portuguese), Chatterbox-Turbo (350M parameters optimized for voice agents with a single-step distilled decoder achieving ~200ms time-to-first-speech), and Chatterbox-Nano (110M parameters running 3x faster than realtime on 8 CPU cores for edge deployment). Unique among open-source TTS systems, Chatterbox introduces emotion exaggeration control — adjusting intensity from monotone to dramatically expressive via a single parameter — and native paralinguistic tagging where tokens like [laugh], [cough], [chuckle], and [gasp] inject natural vocal reactions inline without post-processing. The alignment-informed inference pipeline eliminates hallucinations and repetition artifacts common in autoregressive TTS. Built-in PerTh neural watermarking embeds imperceptible forensic identifiers in generated audio for provenance tracking. Trained on 500,000 hours of cleaned speech data across all supported languages. Voice conversion scripts enable transforming existing audio into any cloned voice. Deploy via pip install with PyTorch, serve through Gradio interfaces or custom FastAPI endpoints, and expose via HTTP streaming or WebSocket for sub-200ms conversational applications. Running on a dedicated VPS on RepoCloud with guaranteed CPU, RAM, and SSD, full root SSH access, and a browser serial console. MIT licensed.

Consolidate your entire digital library into a self-hosted reading sanctuary with BookOrbit, an open-source media platform that organizes ebooks, audiobooks, comic archives, and research PDFs within a personal private cloud. Readers can enjoy multi-format media directly through responsive browser readers that support EPUB, CBZ, and M4B formats without external browser extensions. The platform synchronizes reading bookmarks, highlights, and completion statuses across web interfaces, Kobo ereaders, and KOReader devices. An automated metadata enrichment engine queries fourteen upstream bibliographic indexes, downloading cover artwork, chapter breakdowns, author biographies, and series taxonomies. The staging workspace allows librarians to inspect incoming files, review suggested metadata revisions, and embed updated tags directly into physical storage archives before shelving. Custom reading dashboards track annual book goals, calculate daily streaks, and plot reading habit radar charts. Multi-user configurations support distinct per-user collections, granular content permissions, and single sign-on authentication through OpenID Connect providers. Power users can export private OPDS feeds to external mobile reading applications or push books directly to connected Kindle hardware. Running on a dedicated VPS on RepoCloud with guaranteed CPU, RAM, and SSD, full root SSH access, and a browser serial console. GNU AGPL v3.0 licensed.

Spotify economics without the subscription or catalog gaps: Navidrome, the reference self-hosted music server, streams your own FLAC, MP3, and ALAC collection from a single Go binary with a React/Material UI web player. Its Subsonic/OpenSubsonic API compatibility is the superpower: 50+ existing clients work out of the box, from Symfonium and DSub on Android to Feishin and Sonixd on desktop, plus Android Auto, CarPlay, and Android TV apps. Transcoding is server-managed and FFmpeg-backed - FLAC direct-plays at home and downsamples to MP3, AAC, or Opus over mobile bandwidth, with the OpenSubsonic transcoding extension letting clients declare capabilities and receive per-track direct-play or transcode decisions automatically. Multi-user support gives every account its own play counts, favorites, ratings, and playlists, and multi-library support scopes different collections to different users. The feature list covers serious listening: Last.fm and ListenBrainz scrobbling, artist bios and images, embedded and external lyrics, audiobook bookmarks, saved play queues that resume on another device, internet radio, jukebox mode, and M3U playlist auto-import kept in sync with your folder. Resource usage is famously low - it runs happily on a Raspberry Pi and scales to six-figure track counts.

When a music library is full of wrong tags, missing covers, and blank ratings, Tagr lets you fix the whole collection from any browser, on desktop or phone. Three roles divide the work on your own server. Admins mount a music folder, press scan, and get a searchable table of every track, with sortable and filterable columns, more than 40 editable fields, and room for custom tags. Taggers make the changes: edit fields inline, swap album art, and match one song or a whole selection against MusicBrainz, ticking which fields to overwrite. Mistakes are cheap, because every edit is stored in a persistent change history with a revert action, viewable per song or per folder. Rule-based smart playlists, such as every track with an empty comment, show what still needs attention. Listeners browse and play without edit rights, using the waveform player, a Recently played view, and optional ListenBrainz scrobbling that each person connects to their own account. Supported formats include MP3, FLAC, WAV, AAC, OGG, M4A, M4B, WMA, AIFF, and Opus, with changes written straight back into the files. Swipe-friendly panels, an installable PWA, and five interface languages make the phone a comfortable place to work. 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.

LinkSift is a self-hosted media downloader that turns pasted links into finished MP4 or MP3 files through a calm browser queue you run on your own server. Paste one URL or a whole batch and it first inspects the metadata, showing titles and available qualities before a single byte of media moves. You pick MP4 or MP3, choose the video height the source actually offers, and can even request both outputs from the same link, in which case the audio is extracted from the downloaded video instead of fetched twice. Every job reports its phase, percentage, downloaded bytes, speed, and ETA live, including real encoding progress while the audio is being packaged. Reloading or reopening the tab restores the queue exactly where it was, because a running server job is never re-submitted or cancelled by a closed browser. Finished files save through the normal browser download flow, or into a folder you pick in Chromium-based browsers. A configurable concurrency limit keeps a long batch from saturating your connection, and light, dark, and system themes keep long queues easy on the eyes. There is deliberately no account system or telemetry, so the queue stays a private tool on your network. Running on a dedicated VPS on RepoCloud with guaranteed CPU, RAM, and SSD, full root SSH access, and a browser serial console. MIT licensed.

Stop wrestling with dead scrapers and scattered reading folders by orchestrating your digital library through Bindery, an automated ebook and audiobook management platform that tracks favorite authors, monitors release schedules, coordinates indexer searches, and organizes media collections. Users can configure monitoring profiles to automatically track author bibliographies across OpenLibrary, Google Books, and Hardcover without fear of dead endpoints breaking their library. Dual-format tracking allows each book entry to hold separate digital ebook and multi-part audiobook files in distinct directory roots. The integrated search engine evaluates Newznab and Torznab indexers in parallel, scoring releases by edition tags, audio bitrate, narrator names, and release year to grab the highest quality versions. Download jobs pass seamlessly to background clients like SABnzbd, NZBGet, Transmission, or Deluge, moving completed files into custom library folder structures with hardlinks. An interactive release calendar highlights upcoming publications, while the manual metadata editor locks custom titles, genres, and series ordering against upstream refreshes. Readers can browse curated collections directly through the built-in OPDS feed on mobile reading apps or export organized media straight into Calibre and Audiobookshelf. Running on a dedicated VPS on RepoCloud with guaranteed CPU, RAM, and SSD, full root SSH access, and a browser serial console. MIT licensed.

Listening to music across Spotify on your commute, Jellyfin at your desk, and Sonos in the living room usually leaves your play history fragmented across walled gardens, which is why multi-scrobbler unifies over 30 music sources and routes every track to your chosen scrobble backends automatically. Instead of relying on brittle browser plugins per device, you configure active polling or ingress webhooks once for sources including Plex, Navidrome, Subsonic, Apple Music, Deezer, YouTube Music, MPD, VLC, Chromecast, and Music Assistant. Every detected play passes through an intelligent pipeline that verifies listen duration thresholds, cleans messy track titles or featured artist tags using regex rules, and enriches missing metadata against the MusicBrainz, Spotify, or Rocksky catalogs. Before forwarding a play to Last.fm, ListenBrainz, Libre.fm, Maloja, Koito, teal.fm, Discord Now Playing, or Tuneshine, the engine checks temporal and title similarity scores to block duplicate scrobbles. When a downstream service goes offline, failed scrobbles enter a persistent retry queue so you never lose listening history, while Gotify, Ntfy, and Apprise webhooks alert you to source issues. Running on a dedicated VPS on RepoCloud with guaranteed CPU, RAM, and SSD, full root SSH access, and a browser serial console. MIT licensed.

"The easiest music streaming server available" is mStream's own billing, and the claim holds up: a lightweight Node.js app that turns a folder of audio files into a private streaming service in minutes, no external database required. Its filesystem-based design is the clever part - the API mirrors your folder structure, so you can browse and play music immediately, before any library scan finishes, and your organization on disk is your organization in the app. It streams flac, mp3, wav, ogg, opus, aac, and m4a, which matters to the audiophile crowd: FLAC plays uncompressed, bit-perfect, with gapless playback for live albums and continuous mixes. The web player runs anywhere a browser does and packs personality - a Milkdrop-style visualizer (Butterchurn), playlist sharing via links, and drag-and-drop uploads straight through the file explorer. Native iOS and Android apps add the feature streaming subscriptions can't match: sync your collection to your phone for true offline playback of music you own. Multi-user support assigns separate directories and permissions per account. Resource usage is famously light - mStream is tested on multi-terabyte libraries and runs happily on a Raspberry Pi, so a small RepoCloud instance serves a lifetime's collection. GPL-licensed, with zero listening-habit telemetry.

AudioMuse-AI transforms your self-hosted music collection into an intelligent sonic discovery universe that generates groove-aware playlists without relying on tag metadata or external streaming APIs. The platform scans audio files directly to extract acoustic properties, rhythmic cadence, harmonic density, and lyrical themes across seventy-two supported languages. Listeners can explore their library through a dynamic two-dimensional interactive music map, discover hidden sonic connections between distant genres, or chart smooth transitional song paths between two completely different tracks. Custom prompt-based querying lets users request niche vibes like late-night lo-fi piano, while automated clustering groups acoustically similar tracks into fresh, genre-defying collections. With Song Alchemy, curators fine-tune playlist moods by interactively adding or subtracting acoustic traits before exporting clean tracklists back to Navidrome, Jellyfin, LMS, Lyrion, Emby, or Plex. A built-in acoustic fingerprinting engine even identifies mystery tracks from twenty-second audio snippets, deduplicating songs across multi-server topologies so every file is processed only once. 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.