Kavita

Website: kavitareader.com

Kavita is a free and open-source reading server that you host yourself. Instead of curating a database by hand, you point it at folders of comics, manga and e-books and it scans them into a browsable, searchable catalogue with covers, series grouping, reading progress and readers built into the web interface. It is written in C# with an Angular front end, and it is the fastest-moving of the self-hosted reading servers.

Key Features

• Mixed libraries: Manga, comic, EPUB, PDF and image-folder libraries in one server, each with its own scanning rules and reader
• Wide format support: CBZ, CBR, CB7, ZIP, RAR/RAR5, 7-Zip, loose image folders, EPUB and PDF
• Responsive web readers: Manga and webtoon modes, continuous reading, double-page spreads and virtual pages, plus a dedicated EPUB reader with adjustable fonts, margins, line height and themes
• Reading progress and annotations: Position is remembered per user across devices, and EPUB highlights and annotations are stored in the library
• Multi-user with age restrictions: Role-based accounts, per-library access and age gates — usable as a family server — with OIDC login supported
• OPDS and OPDS-PS: On by default; every user gets their own OPDS URL and auth key, so Mihon/Tachiyomi, KOReader, Panels and similar readers can browse the library and stream pages from the server
• Organisation: Collections, reading lists with CBL import, Want to Read, bookmarks and smart filters, with a dashboard you can rearrange
• Metadata and search: Filtering and full-text search across the whole library, with external metadata, reviews, ratings and recommendations available through the optional Kavita+ subscription
• Localisation and themes: Community translations via Weblate and a theme system with an official theme repository

Licence

GPL-3.0 — the server itself is free and open source, with no account and no telemetry. Two honest caveats. Kavita+ is an optional paid monthly subscription, node-locked to your instance, that unlocks internet-facing extras such as external metadata fetching, scrobbling to Hardcover, AniList and MyAnimeList, and external ratings and recommendations — the base server never loses features to it, but a workflow built around those extras carries a bill. Scanning very large collections can also be slow on the first pass. If you would rather avoid a subscription, Komga (MIT) covers comics and manga with metadata pulls from ComicVine, AniList and MangaDex included, and Calibre-Web (GPL-3.0) is a free web front end for an existing Calibre library.