AskPeppiBot

What it does

askpeppi is a local business directory. Each listing has an assistant that answers questions from visitors: opening hours, whether the kitchen does a vegan option, whether you can book. AskPeppiBot reads the business’s own website so those answers describe the business as it actually is, rather than as a language model imagines it.

We do not train models on what we read. The text is used to answer questions about the business it came from, on that business’s own listing page, and nowhere else.

When it visits

• Only when a person opens that business’s page on askpeppi, or asks its assistant a question. We do not crawl continuously and we do not walk the web looking for sites.
• At most 8 pages per site per visit, chosen for the pages a customer would ask about: the homepage, the menu, hours, contact, about, booking.
• At most one request per second to any one site, and at most one visit to the same site every six hours.
• Once every 30 days per business after a successful read, unless somebody asks a question we cannot answer from what we already have.

We honour robots.txt, including Crawl-delay and Content-Signal. A site that declares use=no is not read at all. If your robots.txt cannot be fetched (a 5xx, a timeout, a 429), we treat that as a refusal and stop, rather than assuming permission.

How to block it

Add this to your robots.txt. It takes effect on our next visit, and within a day at the latest.

User-agent: askpeppibot
Disallow: /

To slow us down instead of stopping us, use Crawl-delay. We respect it up to ten seconds; above that we stop reading your site rather than hold a visitor’s question open waiting.

User-agent: askpeppibot
Crawl-delay: 5

Getting your listing removed

Blocking the crawler stops us reading your site; it does not remove your business from the directory. If you want the listing itself taken down, or you would rather claim it and answer in your own words, write to us from the contact page and we will deal with it by hand.