ComicOPDS is a lightweight OPDS server for a collection of CBZ comics. Point it at a folder of comic archives and it reads the ComicInfo.xml metadata already stored inside each file — series, writer, publisher, year — to build a browsable catalogue. There is no bundled reader: the server publishes the library over OPDS and any compatible client, such as Panels, KyBook 3, Marvin or Chunky on iOS, browses, downloads and page-streams it.
This entry covers the community fork maintained by Nadiar, which keeps Frederik Baerentsen’s original application and adds OPDS 2.0 support, a background scanner, a user administration interface and trusted proxy handling on top of it.
The server negotiates the format from the reader’s Accept header, so an older client keeps getting Atom XML (OPDS 1.2) while a newer one is served JSON (OPDS 2.0) without a second configuration. Page streaming follows the same split: OPDS PSE 1.1 for 1.2 feeds, and Readium Web Publication Manifest manifests in the DiViNa profile for 2.0, which is the recommended way to stream image-based publications such as comics and manga.
X-Forwarded-For and X-Real-IP, so per-client behaviour stays correct behind a reverse proxy.A Docker image and a compose.yaml are the documented deployment route: mount the comic directory into the container, expose the port, and the web dashboard reports what was indexed. The application itself is Python (FastAPI), and the repository also ships a plain requirements.txt for running it outside Docker.
For a full comic and ebook server with its own browser reader, Kavita is the usual pick; Calibre-Web covers the same ground for EPUB collections. ComicOPDS is the narrower tool: an OPDS feed for a CBZ folder, with the reading left to the client you already use.