Update! After a few hours of struggling I managed to set up Read the docs for Tildee, it should help using the library significantly.

After getting some inspiration from TAPS I thought that maybe I try to work on something vaguely similar on my own. And after… some? hours of coding today I came up with this: tildee.py (source)
It's a wrapper for the Tildes Public/Web API that is already used by the site internally to make it work. The obvious problem with that is that it will at one point break when this unstable API is changed. It can do basically all things a normal user can do with the notable exception of applying comment labels (because I haven't gotten around to that yet).

Example of usage for a DM reply bot (result):

import sys
from tildee import TildesClient
import time

# Initialize client and log in, 2FA isn't supported yet and will probably break in horrible ways
t = TildesClient("username", "password", base_url="https://localhost:4443", verify_ssl=False)

while True:
    # Retrieve the "unread messages" page and get a list of the conversations there
    unread_message_ids = t.fetch_unread_message_ids()
    for mid in unread_message_ids:
        # Access the conversation history page; this also clears the "unread" flag 
        conversation = t.fetch_conversation(mid)
        # Get the text of the last message
        text = conversation.entries[-1].content_html
        # Abort if it's from the current user (I don't think this could actually happen)
        if conversation.entries[-1].author == t.username:
            break
        print(f"Found a message by {conversation.entries[-1].author}")
        # If the message contains a reference, reply in kind
        if "hello there" in text.lower():
            print("Replying…")
            t.create_message(mid, f"General {conversation.entries[-1].author}! You are a bold one.")
        # Delay before processing next unread message
        time.sleep(3)
    # Delay before next unread check
    time.sleep(60)

This has a lot of potential. Haven't yet figured out potential for what, but I'll take what I can get.
I'd be really grateful if someone with a little more experience than me (that's not exactly a high bar :P) could give me some pointers on the project's structure and the "API design", hence the ask tag. Other creative ideas for what to use this for are appreciated, too.

Hi,

There's a certain site which hosts media files and has a player that depends on a lot of third-party resources to play, while browsers have native support for those file types. Those 3rd-party resources are often blocked by ad blockers and I have no desire to white-list them. I would like to extract the direct link to the media file and make it playable on my custom web page.

The link to the media file is present in the page source of each page, always on the same line. It's not anchored in HTML but present in the JavaScript for the player, like so:

    $(document).ready(function(){
      $("#jquery_jplayer_1").jPlayer({
        ready: function () {
          $(this).jPlayer("setMedia", {
            [ext]: "https://[domain]/[filename.ext]"
          });
        },

In this example it's on line #5. [ext] = the file extension.

I want to build the following:

• A web page with a form with a single input field meant to receive links from that specific file host
• [Something] that extracts the file link from the source of the host's page
• Present the linked file as playable in an embedded native player

So far I've managed to create a form with an input box and a submit button, but it doesn't do anything yet. What is the best way to build the actual functionality? I know HTML/CSS. I have some rudimentary understanding of JavaScript/jQuery and Python3, so those would be my preferred tools.

For those worried about piracy: The files in question are not copyrighted and I'm not looking to make copies. I just want to make them playable. This is for personal use.

Thank you for reading this far. Any and all advice is welcome!