KanartBot

If you found this page in your server logs, this is who we are and how to make us stop.

User-agent: KanartBot
Disallow: /

Who we are

Kanart is an app that gathers exhibitions in one place — museums, galleries, antique dealers, auction previews, pop-ups — so people stop missing shows they would have liked to see. We are launching in Singapore. It is built by two people, not a company with a crawling department.

KanartBot is the program that reads venues' own "what's on" pages so a listing in the app says what the venue says. We would rather send you visitors than replace you: every listing links to your own page for the show, and that link is the point.

What it reads

Only public pages — the same ones anybody can open without signing in. KanartBot never logs in, never submits a form, never buys a ticket, and never touches anything behind a password.

About each exhibition

• Title and subtitle
• Start and end dates, how precise those dates are, and the dates as your page words them
• Which venue it is at, or where it is if it is somewhere else
• Admission and price, plus your own wording about who pays what
• A link to your page for the show, and a ticket link if you publish one
• Artist names and any category labels you use
• A short excerpt of the description — never the whole thing

About each venue

• Name, and other names it is known by
• Address, postal code, neighbourhood and map coordinates
• Opening hours, and any note you publish about exceptions
• Your website, time zone, and general admission notes

Pictures: we link, we do not copy

We do not download, store or re-host your images. We record the address of an image you already publish, and the app loads it from your server, with a credit and a link back beside it. If you delete or move a file, it disappears from Kanart the same day — which is the behaviour we want, because it leaves you in control.

We also record where on the page each picture sat. Images that read as individual artworks rather than promotion for the show are noted and never displayed. If you would rather we showed none of your images at all, say so and we will switch them off for every listing you have, without touching the listings themselves.

What it does not read

• No personal data. We are not interested in your visitors, your staff or your mailing list.
• No full descriptions, no press releases reproduced in full, no essays.
• Nothing behind a login, a paywall or a form.
• No page your robots.txt disallows — see below.

How often it visits

• At most once a day, and usually less.
• One request at a time per site, with a gap of at least two seconds between them. We are never more than one visitor.
• We ask whether anything changed. Every request carries the ETag or Last-Modified you gave us last time, so an unchanged page costs you a 304 Not Modified and no content at all.
• We slow down if you say so. A Crawl-delay aimed at us is obeyed, and we back off when a server starts refusing connections.

You will see us identify ourselves on every request as:

KanartBot/1.0 (+https://kanart.app/bot)

How to stop us

We read robots.txt before anything else, and we obey it. To refuse us everything:

User-agent: KanartBot
Disallow: /

To let us read your exhibition pages but nothing else:

User-agent: KanartBot
Allow: /exhibitions/
Disallow: /

To keep one area private and leave the rest as it is:

User-agent: KanartBot
Disallow: /private/
Disallow: /admin/

A blanket rule works too: if your file says User-agent: * and Disallow: /, we treat that as meant for us and stay away. A rule that names KanartBot takes precedence over the blanket one, in either direction — so you can single us out to block us, or single us out to let us in.

We actually check

This is not a statement of intent. Every request the crawler makes is checked against your robots.txt before a connection is opened, and a refused address is never requested at all. That behaviour is covered by automated tests that fail the build if the check is removed — both for a rule naming KanartBot and for a blanket User-agent: * rule.

Where your file is genuinely ambiguous, we resolve it against ourselves and do not fetch. There are pages we are fairly sure we are allowed to read that we leave alone for exactly that reason.

Talk to a person

hello@kanart.app reaches one of the two of us. Things worth writing about:

• Stop crawling us — we will, and you do not have to give a reason.
• Something in our listing is wrong. We would rather hear it from you than be wrong in public.
• Do not use our images. One message switches them off everywhere.
• We would rather send you a feed. Gladly — a feed you control beats us reading your pages, for both of us.

Last updated 17 September 2026.