Project of new free drm downloader

[CrocodileMachine]

Hi,
I want to create a new type of downloader that requires the user to manually enter the license URL, the MPD URL, and any tokens, so that there's no need to log into an account, keeping it more neutral and reliable for the user. I'm working on browser automation via Selenium, but that would require passwords, use a huge amount of memory, and seems disproportionate. Supported sites would be limited at first, but depending on users and suggestions, I could grow the list! For the downloader, I plan to incorporate yt_dlp (the project is in Python) solely to parse MPD (and possibly M3U8) manifests.
Does this sound interesting to you?
What do you think would be necessary to add to it?

[RCGn]

Sounds like a plan. I am not going to question how and what.

Therefore, a PoC would be great good luck!

[CWTV]

Manually entering a token sounds like fun.....not. Python can already do all this with scripting, so whilst your desire to keeping it neutral and reliable may have some merit - it just sounds like too much work for the user to interact with.

[Asdfgf]

Manually entering a token sounds like fun.....not. Python can already do all this with scripting, so whilst your desire to keeping it neutral and reliable may have some merit - it just sounds like too much work for the user to interact with.

100% agree

[Obo]

Already exists, eg https://forum.videohelp.com/threads/415503-%5Brelease%5D-AllHellGui or Angela's profile page.

And btw. I think the much smarter idea was to build a browser extension that intercepts collectiing mpd, license url and performing the licensee request. Like https://forum.videohelp.com/threads/416316-%5BRelease%5D-WidevineProxy2-Extension-Bypa...s-Lic-wrapping

But for learning purposes such a project is now doubt useful. (I've also created such a GUI, but since @larley released WidevineProxy2 I haven't used it anymore.)

[CrocodileMachine]

Hi.
I understand that having to manually specify tokens (when they are required) can be annoying and repetitive. However, since I am aiming for a command-line downloader, this information would be passed via command-line arguments. Furthermore, the PSSH could potentially be extracted from the MPD file. Consequently, for most sites, you would only need to provide the MPD URL, the license URL, and—if necessary—a token. Naturally, more complex sites might require cookies or multiple tokens, but the goal is to save time while remaining in control and aware of what is happening at each step. Also, as you mentioned, there are already excellent tools available—such as `widevineproxy2` by larley—so I am leaning toward a Python-based solution. In this approach, a user who already has their browser open could use the web developer tools to extract the MPD URL, the license URL, and, if needed, a token. Moreover, the additional effort involved in installing an extension and passing parameters to a command-line tool is actually quite minimal.
What do you think?
Do you have any other suggestions?
I’ve considered automating the browser via Selenium to retrieve the license URLs, but that might be a bit overkill, don't you think?

[Obo]

I'm in no position to tell you to not implement your idea. You start and you'll see how far you get.

As for yt-dlp - you mentioned it in your start post: have a look at https://github.com/aarubui/yt-dlp-mp4decrypt which is a plugin that integrates into yt-dlp.

[CrocodileMachine]

I was already familiar with this plugin, though not this specific version. It looks cool, but it’s essentially a standalone solution in its own right; integrating it into my project wouldn't really add anything new. That said, I think what I’m planning to build might address the issue of sites that this plugin doesn't support, since my tool would allow for manual input of information—we’ll see! Thanks for the tip.
Please feel free to keep sharing your suggestions and feedback!
Also, are there any specific sites you would like my downloader to support?

[CrocodileMachine]

Could someone point me in the right direction on how to use a Python script to retrieve the MPD URL, the license URL, and any necessary tokens?
(I’m exploring ways to further automate the process, though I haven’t made a final decision yet.)

[phased]

Could someone point me in the right direction on how to use a Python script to retrieve the MPD URL, the license URL, and any necessary tokens?
(I’m exploring ways to further automate the process, though I haven’t made a final decision yet.)

When you first posted I wondered how it would pan out. Now you are asking "how to use a Python script to ... whatever". You clearly know little while ignoring the help you've already been directed to by Obo. "though I haven’t made a final decision yet" - Course you have - and we both have known all along!

[CrocodileMachine]

That’s your opinion, but actually I haven’t made a final decision regarding my project yet.
It’s not that I’ve ignored Obo’s help—I’ve already explored that avenue and found that the plugin for yt-dlp doesn’t work for certain sites, which is why I wanted to create my own solution.
Actually, I didn’t express myself very well; I wasn’t asking for a Python script to do everything. Rather, I was wondering about the feasibility of retrieving license URLs and tokens via Python scripting—at a more advanced level than what I’m currently doing (which is just using the video URL). After all, I don’t know everything!
Sorry, if I come across as too abrupt; English isn’t my primary language.

[larley]

why ai

[CrocodileMachine]

What do you mean ?

[YmFzZTY0]

Does this sound interesting to you?

No it doesnt - your trying to reinvent the wheel, just use unshackle

[CrocodileMachine]

concerning the elaboration of the command line options.
I hesitate between this two syntaxes :
<name> <site_name> <download_mode> <url_mpd> <url_license> <TOKEN> <other_options>
examples : TV tf1 unic "https://myurl/mpd.mpd" "https://myurl/of/license" "TOKEN_VALUE" -T "test" -M mp4
With that I could generate detailled help for what is necessary for each site avec create posisional argument for earch site/command. But is verbose.

Or second option :
<name> <url_mpd> <curl_command_of_license_request> <other_options>
examples : TV "https://myurl/mpd.mpd" "curl.exe -H dgfdjhfgsdj -H dhgfjshdgfjs --data-raw "CAESJHFSDHFDSKHKFDSH" " -T "test" -M mp4
But the help for this syntax will be lessard detailled and the choice of teh service will be deduce from the mpd_url

what's option doc you prefer ?
I think that's the second, but, the first offer for me a better visibility, but it's may be not necessary?

[CWTV]

Generally the way things work around here fall into 2 categories:

1. those who have some idea what they are doing in this space build/release something if they choose too. People then use it, give feedback blah, blah and away it goes.
2. those that know very little, come here ask stupid questions, most likely get stupid answers unless you prove you are genuinely trying to learn and/or use what's already available here.

You seem to fall into a totally new category. Come here knowing nothing, only grandiose ideas and expect others to help you build it. Hate to break it to you, it ain't gonna work that way.

Do some learning. Try out the tools here already. Then if you feel like doing something of your own, great build it first and then release it for people to try out. But it needs to be better or at least as good as what people are already using - otherwise why will they bother ?

[Spectrumhsm]

WidevineProxy2 is my everyday workhorse.
It works perfectly for my sites.
I just added command options:
-M format=mkv -mt --save-name " "
--

Wish it would pull video name instead of me adding manually.

[larley]

This for example is a thread where you can see the young clueless me
I learned all this as well at some point

https://forum.videohelp.com/threads/413795-Extract-Keys-from-License-Server-response

[CrocodileMachine]

thank's for your post larley.
I want to thank you for your incredible work with your extension, your scripts, and your responses in this forum.