Manga Downloader

Manga Downloader is a command line tool that downloads manga chapters from MangaDex and 80+ other sites and packs them into CBZ files. The files are ready to read on your e-reader or in your favorite comic app.

Point it at a manga's index page, tell it which chapters you want, and it does the rest:

manga-downloader https://mangadex.org/title/a1c7c817-4e59-43b7-9365-09675a149a6f/one-piece 1-10

This writes One Piece 0001 - Romance Dawn.cbz and nine more files to the current folder.

• A single binary: no runtime, no dependencies and no config file. Builds are available for Linux, macOS and Windows, plus Docker images.
• 84 supported sites, from big aggregators to small scanlation groups.
• Chapter ranges such as 1,3,5-10, so you download only what's missing.
• E-reader friendly output: AVIF pages are converted to JPEG automatically, so they don't show up blank on Kobo, KOReader or Calibre.
• Flexible output: one CBZ per chapter, a single bundled CBZ (--bundle), or plain image folders (--format raw), named from a template.
• Language and scanlation group filters for sites that host several versions of a chapter.
• Concurrent downloads, tunable per chapter and per page, with retries.
• JavaScript and Cloudflare-protected sites are handled by driving a local Chromium browser, which clicks "Verify you are human" checkboxes for you.

Other sites built on the same platforms often work too, even when they aren't listed. If one doesn't, open an issue or send a pull request.

Download the archive for your system from the releases page, extract it and run the binary:

./manga-downloader --help

To run it from any folder, move the binary to a folder in your PATH, such as /usr/local/bin on Linux and macOS.

docker run --rm -it -v "$PWD:/downloads" elboletaire/manga-downloader [url] [chapters]

Files are written to the mounted /downloads folder. They're owned by USER_ID/GROUP_ID, which both default to 1000. If your user has a different ID, pass it with -e USER_ID=$(id -u) -e GROUP_ID=$(id -g).

The default image has no browser. For sites marked with *, use the :browser image, which ships Chromium and a virtual display:

docker run --rm -it -v "$PWD:/downloads" elboletaire/manga-downloader:browser [url] [chapters]

docker run --rm -it -e DISPLAY=$DISPLAY -v /tmp/.X11-unix:/tmp/.X11-unix \
    -v "$PWD:/downloads" elboletaire/manga-downloader:browser --browser-visible [url] [chapters]

go install github.com/elboletaire/manga-downloader@latest

manga-downloader [flags] [url] [chapters]

• url is the manga's index page, the page that lists all its chapters. Don't use the URL of a single chapter.
• chapters is a comma-separated list of numbers and ranges, such as 1,3,5-10. If you leave it out, the tool asks whether to download every chapter. The answer defaults to "no".

Arguments aren't positional, so manga-downloader 1-50 [url] works too.

Some sites, such as MangaDex, list each chapter once per language. Use --language to download only one of them:

manga-downloader --language es https://mangadex.org/title/a1c7c817-4e59-43b7-9365-09675a149a6f/one-piece 1-10

Sites that host several scanlation groups' versions of the same chapter download the group with the most chapters, and list the other groups. Use --scanlator <name> to pick a group, or --scanlator all to download every version. With all, the group name is added to each file name.

# all chapters in a single CBZ, one folder per chapter inside it
manga-downloader --bundle [url] 1-8

# plain image folders instead of CBZ files
manga-downloader --format raw [url] 1-8

Most e-readers can't display AVIF, so a chapter with AVIF pages looks empty on the device even though the download succeeded. AVIF pages are converted to JPEG by default. WebP renders fine in phone apps but is unreliable on e-ink, so converting it is opt-in:

manga-downloader --convert-images avif,webp [url] 1-10   # strict e-ink setup
manga-downloader --convert-images none [url] 1-10        # keep pages untouched

JPEG, PNG and GIF pages are never re-encoded. If a page fails to convert, it's kept as it was and a warning is shown, so the rest of the chapter isn't lost.

Scanlation groups often reserve a page or two of every chapter for their own credits or ads. --skip-pages drops them, using the same range syntax as the chapter argument, applied to each chapter's pages:

manga-downloader --skip-pages 2 [url] 1100-1110       # just the credits page
manga-downloader --skip-pages 2,5-7 [url] 1100-1110   # several at once

Pages count from 1, and negative positions count backwards from the end of the chapter: -1 is its last page, -3--1 the last three. They're resolved per chapter, so -1 means each chapter's own last page however long it is:

manga-downloader --skip-pages 2,-1 [url] 1100-1110    # one at each end

The pages are dropped before they're downloaded, so they cost no time or bandwidth, and a page that's broken on the site's end can't fail the chapter around it — which is the only way to download a chapter whose page 404s.

The archive is still numbered 000, 001, … in order, with no holes left by the skipped pages. A position the chapter doesn't have is ignored.

Use --filename-template to change file names. It takes a Go text/template string, and the default is:

{{.Series}} {{.Number}} - {{.Title}}{{if gt .Version 1}} v{{.Version}}{{end}}

{{.Version}} is a counter that's added when two files would have the same name.

Run manga-downloader help to list them in your terminal.

You need Go 1.19 or newer:

make install   # download dependencies
make build     # test and build ./manga-downloader