LuvemBooksBot

What it does

When we review a book, LuvemBooksBot reads the published reviews of it that a search turns up: critics' and readers' own write-ups, so our review can say accurately what they concluded and credit them by name, with a link. It reads each page once and keeps the text for 30 days rather than coming back for it. Where a blog's page loads but its text can't be read from the page, it may ask the site's public WordPress API for the same post.

It also reads a small number of public pages (reading lists, “best books” round-ups, genre and author resource pages) to spot broken outbound links to books. When a page we'd genuinely fit links to a title whose destination has gone dead, we may email the page's owner to flag the dead link and, where relevant, offer one of our own pages as a replacement.

When we've published a page about an author's book, it may also read a few pages of the author's own website: to confirm the site is theirs and mentions the book, to find the contact address they publish there, and later to see whether they've linked to us. We use that address only to tell them about our page, once. It doesn't submit forms or access anything behind a login.

For authors we cover, it looks at the official website their Wikidata entry lists (or, when there isn't one, a site with their name that mentions the book we reviewed), and at the news feed that site declares, about every six hours, so we can report their announcements (a new book, a cover, a release date) from their own words and credit them. When we publish a review, it also reads that book's page on their site, found by the link their homepage gives. It reads the homepage, the feed and those book pages, nothing deeper.