SaerBot

If you found SaerBot in your logs, someone used Saer to look at how your website's colours are built: which colours are defined, and where the pages use a raw value instead. It reads what any browser downloads, slowly, and leaves if your robots.txt asks it to.

How it identifies itself

Every request carries this user agent:

SaerBot/0.1.5 (+https://saer.lunarcomputing.org/bot)

The version number changes with each release; the SaerBot token does not.

What it reads

• The HTML of the page it was given and a few pages linked from it on the same site (5 by default, never more than 20).
• The public stylesheets those pages load.
• Your robots.txt, before anything else.

It does not run JavaScript, submit forms, log in, send or store cookies, or follow links to other sites. It does not collect personal data, and it is not a search engine or an AI training crawler.

How politely

• At most one request per second, and slower if your robots.txt sets a Crawl-delay.
• At most 15 MB in total per audit and 4 MB per file, with a 15 second timeout.
• Ports 80 and 443 only. It refuses private, internal and loopback addresses.
• If your robots.txt cannot be read (a server error or no answer), it reads nothing.

How to block it

Add this to your robots.txt. SaerBot checks it on every visit and on every host it fetches from, including the host that serves your CSS:

User-agent: SaerBot
Disallow: /

If your robots.txt has no section for SaerBot, its User-agent: * rules apply.

What happens to the results

Saer runs on the computer of whoever started the audit; the results stay with them. When we use an audit in public, the site's name, addresses, page paths and file names are removed first. A report that names your site is only ever shared with you.

Contact

Questions, or SaerBot misbehaving? Email hello@lunary.app with the time and the user agent from your logs, and we will look into it. SaerBot is made by Lunar Computing, Inc.