RightSignal Legal · Bot information

The RightSignal bot

This page is for website owners and anyone who has seen “RightSignal” in their server logs.

What RightSignal is, and why it reads sites

RightSignal Legal provides provenance and rights evidence for AI and data companies, scraping agencies, law firms and in-house counsel. To do that, RightSignal reads a site to record what the site permits for AI training and similar uses, and keeps a record of what it found and when.

How it identifies itself

Every request carries RightSignal’s own declared User-Agent:

RightSignal/1.0 (+https://rightsignal.legal/bot; rights-signal observation)

The robots.txt product token is RightSignal.

What it reads

RightSignal reads a site’s terms and policy pages to learn what the site permits, including where robots.txt disallows them, and records that it did so. It reads at most 10 such pages in one check. Apart from your signal files, such as robots.txt itself, it reads nothing else against robots.txt.

How RightSignal reads pages

RightSignal may render a site’s home page in a browser when its links are drawn by scripts, to find its policy pages. When it does, it renders the home page once, in a browser that sends RightSignal’s own User-Agent, the same one named on this page. It reads robots.txt for every host the page loads from and keeps to it. It loads no analytics or advertising, makes no bot-challenge or bot-scoring request, and sends no form or other request that would change anything on the site. It clicks, types, accepts and solves nothing. It stops at a 429, a challenge or a block. It loads at most one page and 60 resources, within 90 seconds.

When RightSignal cannot find a site’s terms, an assistant may open up to five of the site’s own pages, under the same User-Agent and the same rules, to look for them. RightSignal may also use a language model to read the terms pages it finds. Only the text of the site’s public policy pages, and of the pages opened to find them, is sent to the model. The model never decides an outcome: every quotation is checked word for word against the page, and its reading can only make RightSignal’s answer more careful, never less.

How it behaves

Keys for signed requests

RightSignal publishes the public keys for its crawler in a key directory at:

https://rightsignal.legal/.well-known/http-message-signatures-directory

Request signing with Web Bot Auth (HTTP message signatures) starts once Cloudflare has verified the crawler. Until then, recognise RightSignal by its User-Agent.

Allowing or blocking it

RightSignal obeys the robots.txt rules for RightSignal and for *. To allow it, you need do nothing more than leave it out of your disallow rules: where robots.txt does not disallow it, it reads within the limits on this page.

To block it from your content pages, add this group instead:

User-agent: RightSignal
Disallow: /

RightSignal will still read your robots.txt to see the group, and it will still read your terms and policy pages as described above, so that it can record what you permit. To slow it down, set a Crawl-delay or answer with 429 or Retry-After. To refuse it altogether, answer its requests with 401 or 403, which it never passes, or write to crawler@rightsignal.legal from an address at your site’s domain.

Contact

Questions, concerns or complaints about the bot: crawler@rightsignal.legal.

The full Crawler and Site Owner Policy is at rightsignal.legal/crawler.