# Perplexity blocked in robots.txt and still in the log | Illia BizTech

> The rule sits in the file and the requests keep arriving. Perplexity runs two robots, and the documentation says something different about each.

Source: https://illia-biztech.de/en/forum/perplexity-ist-in-der-robots-txt-gesperrt-und-steht-weiter-im-log

1. [Home](/en)
2. [Forum](/en/forum)
3. Perplexity is blocked in robots.txt and still turns up in the log

# Perplexity is blocked in robots.txt and still turns up in the log

Weeks with the blocking line in the file, and requests carrying that name keep arriving.

* ##### Asked by  
RestlessHarbor287
* ##### Asked on  
September 13, 2026

We run a specialist shop with roughly three hundred items. In the summer of 2026 we decided not to hand our copy to AI providers, and we put a blocking line for Perplexity into robots.txt.

Several weeks have gone by. The server log still shows requests with Perplexity in the name. Our contractor says we wrote the file wrong. It looks right to me.

#### Answers

![Portrait of Illia Belichenko, technical specialist at Illia BizTech](/_ipx/f_webp,q_75/https://cms.illia-biztech.de/uploads/small_Illia_Belichenko_author_81f88864f0.png)

##### Illia Belichenko · Founder & Technical Lead

✓ Accepted answer September 14, 2026 

The blocking line is probably written correctly. Perplexity runs two robots with different jobs, and they treat robots.txt differently. Each carries its own name and its own address inside the user-agent string, so a filter written for one of them never matches the other.

The first is called PerplexityBot. The documentation describes it as the robot that surfaces and links websites in Perplexity search results, and says in the same paragraph that it does not crawl content for AI foundation models. Blocking that one removes exactly that, the chance to appear there and be linked from it.

The second is called Perplexity-User. It visits a page once somebody has asked a question, and the documentation says it may include a link to that page in the answer. The same page holds the sentence that explains your log. Because a person requested the fetch, it generally ignores robots.txt rules. Your file is not the problem here.

The vendor says a change to the file may take up to twenty-four hours to reach its systems, so a check the same afternoon tells you little. The real question is how visible you want to be in answers like these. Start by looking at which of the two names appears in your own log. The [AI visibility report](https://illia-biztech.de/en/ki-sichtbarkeitsbericht) at Illia BizTech measures which AI answers the shop appears in.

##### 

#### [The report puts a monthly value next to a yearly average](/en/forum/im-report-steht-ein-monatswert-neben-einem-jahresschnitt)

September 11, 2026

[](/en/forum/im-report-steht-ein-monatswert-neben-einem-jahresschnitt)

#### [We never set a radius and the campaign is running in one anyway](/en/forum/wir-haben-nie-einen-umkreis-eingestellt-und-die-kampagne-lauft-trotzdem-in-einem)

September 11, 2026

[](/en/forum/wir-haben-nie-einen-umkreis-eingestellt-und-die-kampagne-lauft-trotzdem-in-einem)

#### [We are being sold a file that supposedly helps Google understand our site](/en/forum/uns-wird-eine-datei-verkauft-mit-der-google-unsere-seite-angeblich-besser-versteht)

September 11, 2026

[](/en/forum/uns-wird-eine-datei-verkauft-mit-der-google-unsere-seite-angeblich-besser-versteht)

#### [The location block in our ad was rejected and Google Ads has no field for it](/en/forum/der-standortblock-in-unserer-anzeige-ist-abgelehnt-und-andern-lasst-er-sich-in-google-ads-nicht)

September 10, 2026

[](/en/forum/der-standortblock-in-unserer-anzeige-ist-abgelehnt-und-andern-lasst-er-sich-in-google-ads-nicht)

#### [One switch on the domain removed our whole shop from the AI features](/en/forum/ein-schalter-auf-der-domain-hat-unseren-gesamten-shop-aus-den-ki-funktionen-genommen)

September 10, 2026

[](/en/forum/ein-schalter-auf-der-domain-hat-unseren-gesamten-shop-aus-den-ki-funktionen-genommen)

**Share** 

**Topics** 
* Perplexity
* robots.txt

A blocking line in robots.txt is always addressed to one name. Perplexity search publishes two of them, standing for two different processes. The first belongs to building the pool that answers are later assembled from. The second belongs to a single question somebody has just asked.

Which of the two turned up in the log decides what the block actually did. Appearing in the results rests on the first name, while the second keeps fetching pages afterwards. The vendor publishes both user-agent lines in full, and one filter line in a log shows which of them arrived.