A couple of things about the video:
1. I disabled the cache and commented out part of the script so that tiles are loaded only from your path.
2. I set Count Parent Tiles to 0 so that parent tiles do not affect the view of the map, and tiles are unloaded when zoom in, and loaded again when zoom out.

No, you don't have to resolve it manually.

How tile loading works:
1. If a tile is in the cache, it is loaded from the cache.
2. If someone has subscribed to OnStartDownloadTile, this event is called, where you can load the tile yourself or do some action, and continue downloading using the built-in downloader. If you loaded a tile yourself, then it is not cached.
3. If the settings indicate that a tile should be loaded from Resources or Streaming Assets, the tile is searched for in these folders.
4. The tile is loaded from the Internet.

So you need to subscribe to OnStartDownloadTile, and download the tile from your folder.

How to get the provider id:

private void OnStartDownloadTile(OnlineMapsTile tile)
{
    // ...
    
    OnlineMapsRasterTile rTile = tile as OnlineMapsRasterTile;
    string provider_id = rTile.mapType.provider.id;

    // ...
}

Tip: do not use the cache folder for your tiles, because this will create two problems:
1. The cache will not load and/or delete these files because they are not in the atlas.
2. If you delete them yourself, then how do you know if this file is not in the atlas, which can create dead records in the atlas.

Where do you save the data in the script you implemented? To Resources folder?
You don't have write access to Resources after build.
Typically, Application.persistentDataPath folder is used for these purposes.
https://docs.unity3d.com/ScriptReferenc … aPath.html

TileDownloader stores tiles in persistentDataPath.

The cache is also not stored in Resources, and the file cache tries to load only the tiles that are present in the atlas.