The quiet race to label synthetic media
Newsrooms and platforms are adding provenance labels, but most images still arrive without them. Readers are left to judge for themselves.
Show readers the same quiet, explainable scores they know from the Lantern extension, on your articles, comments, listings and profiles.
Readers see a likelihood beside the story, its images and the people in the comments. Nothing else on the page changes.
The quiet race to label synthetic media
Newsrooms and platforms are adding provenance labels, but most images still arrive without them. Readers are left to judge for themselves.
Lantern runs in your readers’ browsers, on the content you choose. Your site pays for the checks from its own allowance, so readers never sign in or subscribe.
It’s one small file with no dependencies. It finds the content you marked for scoring, and with auto detection on, article text and large images too.
The reader’s browser sends that content to Lantern with your site id, and Lantern accepts it only from your domain. Anything checked before comes back from the cache at no charge, so a popular page costs you once.
A score appears beside each item, with the same evidence card as the extension. If a check fails or an allowance runs out, nothing appears and the page works as it always did.
You need your site id and access to your page templates. There’s nothing to install on your server.
Talk to our team and we set up your site with its domain, a monthly allowance for each kind of check and a spending limit. Its id then appears in your dashboard under Settings, in Connected website integrations.
Paste the tag into your page template, just before the closing </body> tag, with your site id in data-lantern-site. Pages served over https from your domain and its subdomains work straight away. For another address, such as a staging site, ask us to allow it.
Mark the content to check with data-lantern-type, or add data-lantern-auto="true" to the tag to find article text and large images by itself. Set data-lantern-root to the element that holds your own content, so comments and embeds you don’t control never use your allowance.
Scores appear beside the content within a few seconds. If your site sends a Content-Security-Policy, allow the script and the API in it. When something’s wrong, the browser console says what in one line that starts with [lantern].
<script
src="https://lantern-ai.app/widget.js"
data-lantern-site="SITE_ID"
async
></script>Paste this before the closing </body> tag of your page template, with your site id. It shows scores for the content you mark up.
Your readers see the same scores extension users do, and you decide how much Lantern is allowed to spend.
Scores, confidence and evidence match the Lantern extension exactly, so trust carries over to your pages.
Each category has its own monthly allowance, under a spending cap you agree with us. When a category runs out, its scores stop appearing until the allowance resets, and readers never see an error.
Lantern keeps derived results only. Private items stay scoped to your site and are never shared.
Pricing is negotiated on volume. Contracts set monthly allowances per category, spending limits and overage rates. There’s no unlimited plan.
2 comments
Labels help, but only if people can see why something was flagged.
Great read! Grab 70% off AI photo tools today only, link in my profile.