ManyBooks dates from the early 2000s, when someone looked at the raw Project Gutenberg pile and built a friendlier face over it: genre browsing, short blurbs, recommendations. Much of the stock traces back to those same public-domain scans, rewrapped; the site is independent and ad-supported, with a paid membership layered over a catalog that stays free.
Downloading needs no account and no money: pick a title and take EPUB, PDF, MOBI, or several other formats. The membership mostly buys conveniences — a calmer interface, shelf tools — rather than different books. Everything on offer is English-language.
Quality is the honest caveat. Files are volunteer- and machine-processed reflows of old scans, so typography ranges from clean to eccentric, and when the same title exists at Standard Ebooks that version is nearly always better made. Use ManyBooks to browse and stumble into things you did not know you wanted, use Gutenberg for depth, and look at the first few pages of any download before trusting it on a long read.