SellomBot, Sellom’s crawler

Last updated: 2026-10-15

What SellomBot is

SellomBot is the reading crawler of Sellom, a service run by Enumalon SAS that measures the visibility of brands in AI answers. It reads a few public pages of a website to write the questions of the measurement and to pick up technical signals: robots.txt, llms.txt, structured data.

How to identify it

SellomBot identifies itself by name, with this header:

User-Agent: Mozilla/5.0 (compatible; SellomBot/1.0; +https://sellom.io/bot)

The name to use in robots.txt is SellomBot.

If a site refuses that name (error 403 or 406), SellomBot tries once more presenting itself as a common browser, then reads the site’s other pages the same way, for the same measurement. It does not get around a filter that asks the visitor for a check.

The AI services we query may visit your site with their own crawlers (ChatGPT-User, PerplexityBot…): those visits do not come from SellomBot.

When it visits

SellomBot reads a domain in three cases only: someone runs a test of that domain on sellom.io; a Sellom subscriber prepares or checks a correction on a page of the domain in their dossier; a measurement scheduled by a subscriber reaches its date. Anyone can run the test of a domain: a visit does not mean the site owner asked for it.

SellomBot does not roam the web: it reads the tested domain, its www variant and, at most, the address to which their home page or their llms.txt file redirects. It follows at most three redirects per address. Before entering another host, it reads that host’s robots.txt file.

What it reads

First robots.txt. Then it checks, without reading its content, that the home page does not redirect outside the domain.

It then reads the home page, llms.txt, the sitemap (sitemap.xml) and at most four other public pages, chosen preferably among service, offer, pricing or about pages, whether robots.txt allows them or not. During the measurement it reads the home page and llms.txt again.

Limits for one test:

  • at most 10 pages read, not counting robots.txt and redirects, and 2 when the domain’s questions already exist in the same language;
  • at most 4 simultaneous requests;
  • at most 10 seconds to read robots.txt, redirects included;
  • at most 10 seconds of waiting for a response, at most 800 kB per page;
  • at most 60 seconds for the first reading of the pages, 30 seconds for the re-reading during the measurement.

Judging by their address, it skips files (images, PDFs, styles, scripts) and account, login, cart or legal notice pages. It does not run JavaScript, fill in forms or send cookies back. The text of the pages read is sent to an AI model to describe the business and write the questions (see the privacy policy).

How often

One visit per test, never continuous. SellomBot reads robots.txt again on every test. The Crawl-delay directive is not defined by the RFC 9309 standard: SellomBot does not apply it. Its reading stays within the limits for one test given above.

What SellomBot does with robots.txt

SellomBot reads the site’s robots.txt file before any page and reports it in the report. It interprets it according to the RFC 9309 standard:

  • it uses the group User-agent: SellomBot if there is one, otherwise the group User-agent: *;
  • the most specific rule wins; if two rules are equally specific, Allow wins;
  • only the first 500 KiB are parsed, within 10 seconds, redirects included, once per test.

SellomBot reads the pages on behalf of the Sellom user who requests the measurement. A file that blocks it from the home page, that blocks AI answer crawlers or that asks for ai-input=no (Content-Signal) therefore does not stop the reading: the measurement takes place, and the report states what the file blocks, with the list of AI crawlers concerned and the lines to correct.

Naming SellomBot in robots.txt

To state that you refuse SellomBot on the whole site:

User-agent: SellomBot
Disallow: /

To allow it on the whole site:

User-agent: SellomBot
Allow: /

These lines are read and reported in the measurement report; they do not prevent a reading requested by a Sellom user. A change to robots.txt is visible from the next test. For your site to no longer be read by SellomBot, write to us (see Contact).

Contact

A question about SellomBot, or a visit you do not understand: contact@sellom.io. Please give the domain, the date and the time of the visit, with the time zone.

Publisher

SellomBot is operated by Enumalon SAS (SELLOM by Enumalon), 584 Avenue du Lac, 07350 Cruas, France, RCS Aubenas 101 694 347. See the legal notice.