Skip to content

Servers

Jump to bottom
Proxymiity edited this page May 26, 2021 · 3 revisions

What is MD@H?

MD@H, mdah or simply MangaDex@Home, is a user-hosted, cache server for MangaDex.
The network is made of volunteers that runs their own servers or computers, for the sake of the Horny Jail.
"Basically, your machine will act as a server where a tiny portion of older MangaDex chapters will be stored and when a reader wants to read an older chapter, it will be "fetched" from your machine and served to the reader." (more info on MD's mdah page)

How to download from MD@H?

So, basically, if you want to download from MangaDex, you'll have to use the API to fetch images. A Chapter has a set of pages. You can retrieve the list of pages by calling the get_chapter() function with a chapter ID:

>>> import MangaDexPy
>>> cli = MangaDexPy.MangaDex()
>>> manga = cli.get_manga('b9797c5b-642e-44d9-ac40-8b31b9ae110a')
>>> chapter = manga.get_chapters()[0]
>>> chapter.pages
['1-3a604117709612f29d1a2d21612822f676bcb38a4f378e73fec2c7f1298afac4.jpg',
'2-619df12bbb7eeab3eea0f351e303279cad6aac8aaf07289adec678a571262bff.jpg',
...,
'26-f50267e479b81cfa6d0622400e49dceaafd085703d063ee069e61f1db44f309c.jpg',
'27-5971dd9f63218c00260c5f5e95adcdd3dff002f7f1278216345197a392223854.jpg']

These are the page filenames as they are stored on MD@H clients. To retrieve a page from a client, you'll need to be authenticated with a token, which is included when you request a Network Chapter via Chapter.get_md_network().
You can then get a link with NetworkChapter.pages using the page index (0 being the first page, -1 the last, etc.). (You can also use the .pages_redux property to get lighter pages)

>>> net = chapter.get_md_network()
>>> net.node_url
'https://agrfaez4v18ee.rmhkd98sb0x3y.mangadex.network:443/[........]'
>>> net.pages[0]
'https://agrfaez4v18ee.rmhkd98sb0x3y.mangadex.network:443/Pgx5XF_-KmSobNizAJay9CRTHUaK3nH2VPnM4gFCu7mtib8YThA7EkGlfSTquIFsVGMemzXSTL5lNY4B5sLWClLakIt9qPKJhXD6l8tcUGR30u0TB7YxcdxWvIEMMxCz00g3gGgzxjKdfZAoPJQzz6a4QvygcRR7Rs7X46vNa0J4a8D00081dQPF-WF5Aezd7XejgvRs67x950jcGNl0rnphp8Mmcg/data/7d68dbe3f684dc4c1a68d98179037aa6/1-3a604117709612f29d1a2d21612822f676bcb38a4f378e73fec2c7f1298afac4.jpg'

These URLs will lead you to the page: url test demo

However, you need to take this into consideration when using the pages system:

  • Tokens for these pages expires after ~15 minutes. When these will be expired, you won't be able to browse them. To counter that, you can use Chapter.get_md_network() again. This call will regenerate a new token and assign you a new server
  • If you can't reach the server, consider reporting (see below) before asking a new server, otherwise you'll be likely assigned the same server.
  • Please, for the network's sake, do not spam huge requests to MD@H clients. There is a tool included with this repo to download chapters, which minimizes the network load on MD@H clients by only refreshing one time, and that will automatically report network status, etc.

After downloading from MD@H - Reporting

After downloading a page with MD@H, you should report its status. This allows MangaDex to build anonymous statistics on their servers, and allows them to remove servers that no longer works.

>>> net = chapter.get_md_network()
[... some downloads using requests, for example ...]
>>> net.report('https://agrf[...]98afac4.jpg', success=True, cached=True, req_bytes=120000, req_duration=300)
True

The above example shows that you've successfully downloaded(or 'used' in any means) image with url https://agrf[...]98afac4.jpg. The X-Cache header was set to HIT, you've transferred 120000 bytes in 300ms.
Please see this link for more info on how to report correct stats.

Clone this wiki locally