Selenium + Python automation of Spotify Web App, Dashboard control

I need an expert in Python and Selenium+Chrome automation to create a web dashboard system to automate Spotify's webapp. It must be able to launch a high number (100+) instances, each with different proxies, to control the web app. – Main functions: – Create accounts – Create playlists and add songs via URI to playlists – Stream from playlists, and save/“like” songs * All functions need to use gateway proxies. Gateway proxies change IPs every X minutes automatically. The instances need to check if the gateway proxy's outgoing IP has changed. * Since captchas are used on the web app, I'd like 2captcha to be useable via their API keys, including a notification if the 2captcha balance is low. * When switching to a new account (whether creating accounts or using accounts), a new Chrome/Selenium session is cleared of all cache and cookies, and the system waits until the gateway proxy's outgoing IP has changed. * Gateway proxies have a specific time in minutes before their outgoing IP changes, so the user can enter that timing. The system will need to check frequently for IP changes, to try and protect against a session continuing after the gateway switches to a new IP. * If a gateway proxy switches to an outgoing IP that has already been used for an account, that IP can be used for activity on that specific account, but that IP cannot be used to make another account. * If a gateway proxy switches to an unused IP, it can be either used for an existing account or for creating a new account. *The use of automation via Selenium must be completely 'cloaked' from any automation detection in Chrome. Global user provided data: – List of targeted songs – List of non-targeted songs (only used for playlist creation) Specifics for creating accounts: – While creating accounts, I'd like the current external IP to be noted and associated with that account. Since gateway proxies change IPs every X minutes, there may be multiple IP's per account. – Account parameters are randomized with realistic values – User options – range of time in seconds to wait between each account creation (this is – per instance) – Maximum number of accounts to create Specifics for creating playlists / adding songs - Create a playlist, name can be generated using a spintax string. - User option: % chance to add a targeted song to playlist, % chance to add a non-targeted song to playlist - After playlist is created and songs are added, that playlist url is automatically added to a list of ‘personal’ playlists Specifics for streaming: - A list of target playlists urls are provided, and a list of targeted song URIs (or artist + song titles if a URI isn’t easy to track in web app) - ‘personal’ playlists are also considered, which are playlists created by the instance - A random playlist is picked from the list, and the instance starts at the first song on the playlist. The instance checks on if it’s a targeted or non-targeted song. For a non-targeted song, the instance uses the % chance to skip a non- targeted song, and will either play or skip it, based on the settings below. If it’s a targeted song, it plays the song, using the settings below. After playing, the instance scrolls to the next song and the process repeats all the way down the playlist. If the instance plays all the songs in the playlist and still has time before the gateway proxy changes, it picks another playlist to play from. - User Settings – % chance to “like” a targeted song – % chance to “like” a non-targeted song – range of time (in seconds) to stream a targeted song – range of time (in seconds) to stream a non-targeted song – % chance to fully play a targeted song – % chance to skip a non-targeted song – max number of times an instance can play one specific targeted song (so in a given session, the same song isn’t played too many times)