What is ShapBot?
ShapBot builds part of the web index behind Parallel's search and content extraction APIs. Parallel's crawler documentation says it discovers and indexes websites for those APIs. Developers can request the resulting material through Parallel's web products.
This is index-building traffic rather than a fetch tied to one named end user. Parallel publishes the full user agent as Mozilla/5.0 AppleWebKit/537.36 (KHTML, like Gecko); compatible; ShapBot/0.1.0. It also publishes a current IP list, which gives site operators a stronger check than the ShapBot token by itself.
Allowing ShapBot can make a public page available to Parallel's web APIs. What happens to returned content depends on the application using the API. A crawl does not guarantee placement in any particular answer, and it does not establish that Parallel will cite the page in a public search interface.
Parallel describes search and extraction as the purpose of this crawler. Its documentation does not say that ShapBot gathers foundation-model training data. Blocking it is therefore an opt-out from this web index, not a documented opt-out from model training elsewhere.
Robots behavior needs a cautious reading. The bot metadata marks respectsRobotsTxt as true, while Cloudflare's record says ShapBot does not follow robots.txt. Parallel asks webmasters to allow the crawler in robots.txt but does not make an explicit compliance pledge on its crawler page. Publish a rule to state your preference, then use an enforced control if the distinction matters.
Cloudflare identifies Parallel as the operator and points to the same crawler documentation. No Web Bot Auth signature directory is listed. For a high-confidence allow rule, require both the documented user agent and a source address from Parallel's published list, while allowing for that list to change.
