For publishers and site owners

What Diplet copies, and when

Diplet never crawls the web and never copies a page on its own. A copy of a page is made only when a person who is signed in asks for it, to read it with their group:

• By link. A member adds a link to a meeting or to their own reading room, or shares it with the app, and our server loads that one page once. We keep a plain reader-view copy of its text, so that everyone in the group comments on the same text; the page's scripts, styles and tracking are dropped, and its images are shown from your site, not copied. A link to a PDF is stored as a PDF.
• By browser extension. A member clips the page they are looking at. The extension reads it in their own browser, as they see it, and sends us the text.

The copy is visible only to that person, or to the group they added it to; it is not public. The link back to your site stays with every copy, and a group can suggest a link for its reading list without any copy being made.

Asking us to exclude your site

The owner of a site, or its copyright agent, can ask us to exclude a domain: write to our contact address with the domain and, if you are writing for someone else, on whose behalf. We add the domain to a list that our administrators keep (and every change to it is logged). From then on:

• The domain covers its subdomains: listing example.com also excludes www.example.com and news.example.com.
• No page on it becomes a reading: our server does not fetch it (nor a short link that leads to it), and the browser extension does not clip it. The person who tried is told that the publisher asked not to be copied.
• The link itself is not refused. A group can still keep it on its reading list, and read the page on your site.

Copies made before your request are not removed by it. To have a particular reading removed, send a copyright notice as described in section 6 of the terms; we then also stop the same link or text from being added again.

Signals your pages can send

You do not have to write to us. Diplet honours these hints, which your site can publish for every page it wants kept out:

• A robots meta tag in the page's head: <meta name="robots" content="noarchive">
• An X-Robots-Tag: noarchive HTTP header, which also works for PDFs and other files that have no head.

noarchive, nosnippet and none all count: a reader-view copy shows all of a page's text, which is more than a snippet, and none is a site saying that it wants to be left alone. Other values (noindex, nofollow, max-snippet and so on) are about search results and links, and do not stop a person from keeping a copy of what they read. They can be combined with commas, as usual.

To address Diplet alone, use its name: <meta name="diplet" content="noarchive"> or X-Robots-Tag: diplet: noarchive. Our server fetches pages the way a browser does, for a person, so it has no user agent of its own to match against. When the browser extension clips a page it reads the same robots meta tags in the tab and sends them along, and the server applies the same rule. A page that sends such a hint is not copied, and the person who tried is told why.

robots.txt

Diplet does not read a site's robots.txt. That file tells crawlers which pages they may visit on their own, unasked. Diplet is not a crawler: it loads a page only because a person asked for that page, and the browser extension reads a page the person already has open, in their own browser, where robots.txt has no say at all. It is the same as opening the link in a browser, so a robots.txt rule that keeps crawlers out does not keep a reader out.

If you want your pages kept out of Diplet, use the list above or the hints in the previous section. They are what we check.