>>105997621I have an idea, but I'm unaware of its feasibility. It's a torrent tracker scraper.
1. The user enters the torrent hash.
2. The site uses the hash to query all known, public torrent sites like 1337X, BT4G, TPB, BTDigg, Torrentz2, SolidTorrents, etc. It also queries search engines.
3. It downloads any *.torrent files it finds and extracts the trackers.
4. It does the same as above with magnet links.
5. It combines the trackers into a torrent file.
6. The user downloads the torrent file.
7. If another user enters a previously entered hash, it rescrapes.
8. If no new trackers have been found, the current torrent file gets offered.
9. If new trackers are found, the torrent file gets updated.
The only other features I can think of: a real-time status checker (seeds, peers, availability, last seen complete, etc) that updates during the scraping phase. And a vote (fast, clean ones get positive votes) and/or comment (warnings for slow downloading or malware) system. As far as future expandability goes, a searchable database where users can search for torrents via keywords that shows how many torrents exist for an item. And the ability to sort by seeders, peers, availability, last seen complete, etc. And have torrents from private trackers marked.