naulon – Get paid when AI reads your work

define( 'NAULON_API_KEY', 'nln_live_…' );

Changelog

0.7.3

• New name: naulon – Get paid when AI reads your work. Nothing else changed.

0.7.2

• Setup no longer warns that a verified site “cannot pass” verification because its WordPress address says http. A verified site already passed the https check, often behind a proxy or CDN.

0.7.1

• The plugin’s answer about who is paid for a post or file is never stored by a CDN. Cloudflare caches any address ending in .pdf, so it kept the answer for a tolled file for four hours, and a changed author or wallet was not seen until it expired.

0.7.0

• Charge AI agents for files. Tick PDF, JSON, TXT or XML on your site’s page in the naulon dashboard and the plugin tolls those files in your media library the way it tolls articles. People still download free. A file is credited to the post it is attached to, with that post’s authors and price, and a file on a free post stays free.
• The plugin checks your server can do this before the dashboard offers it. It asks your own site for a small test file and only reports “working” when WordPress itself answered. Diagnostics shows the result, with the exact fix when your server ignores .htaccess (nginx), cannot write it, or a cache in front serves files on its own.
• Saving your key now lights up your site in the naulon dashboard straight away, instead of waiting for the first crawler.

0.6.5

• Co-written posts are credited to every author. Co-Authors Plus, PublishPress Authors and Molongui Authorship are read automatically, with nothing to set up. Each author gets an equal share, paid to the wallet on their own profile. A guest author with no WordPress account is credited too, without a wallet, so your naulon setting for unclaimed author shares decides where their share goes. Before this, only the post’s main author was credited.
• AI agents can no longer read a charged article free through WordPress’s other routes to it. The REST API (/wp-json/wp/v2/, including lists and ?rest_route=), feeds (RSS and Atom) and listing pages such as the home page now give an agent the teaser and a link, and the article itself is where it pays. People, signed-in users and crawlers your rules let read free still get everything.
• A request is no longer treated as your own site because of a Referer, Origin or X-WP-Nonce header. Any client can send those, and each let an agent read every article free.
• New catalog endpoint, /wp-json/naulon/v1/catalog. naulon now reads your list of articles from the plugin itself, with every author under the same id they are paid under, instead of from the public WordPress API.

0.6.4

• Observe mode. When your naulon dashboard is set to observe, AI agents read your posts free while the plugin prices each read and reports it, so you can see the demand before charging anyone. Payments are not taken while observing. Blocked crawlers stay blocked.
• Reports whether a crawler’s stated price (crawler-max-price) covered yours, on charged and observed reads alike.

0.6.3

• Authors’ profiles and the Earnings screen now link to naulon’s withdrawal page. Payments settle to each author’s Circle Gateway balance, not straight into their wallet, and the plugin never said where to withdraw them.
• Corrected the profile and Earnings text that said there was nothing to withdraw.

0.6.2

• The External services section now says what an author approval sends: the author’s email address, their id on this site, and the approver’s email address. It had left this out.
• Settings → Privacy now offers a suggested paragraph listing everything the plugin sends to naulon.

0.6.1

• Diagnostics names a caught impostor even when it read free, calls an allowlisted crawler an allowed crawler rather than a person, and lists the companies whose crawlers’ impostors your site charges.

0.6.0

• Requests that claim to be a known crawler are now checked against that crawler’s published addresses. Once naulon has seen the real crawler on your site, an impostor using its name no longer reads for free.
• New setting on the Content screen: the trusted client IP header, for sites behind a proxy or firewall other than Cloudflare.
• Responses that must not be cached now also carry LiteSpeed Cache’s own no-cache header.
• Diagnostics shows, for each crawler decision, whether its address was checked and what it showed, and adds the crawlers this site checks to the user agents to keep out of your page cache.

0.5.9

• A request that cannot pass the signature checks no longer makes your site fetch naulon’s key list. After a failed fetch the plugin waits ten seconds before trying again, so a slow key server cannot hold up your pages.

0.5.8

• Serves naulon’s own signed origin fetch without charging, so the plugin can run beside a crawler route.

0.5.7

• Opening your licence at /license.xml in a browser now shows the terms instead of downloading a file. Crawlers still receive it as an RSL document.

0.5.6

• The crawler choices you make on naulon.app now apply on this site. A crawler you block there is refused here, one you let read free reads free, and one you charge is charged. Before this they only applied to sites routed through naulon.
• What AI may do, set on naulon.app, is enforced here too. Refusing training turns the training crawlers away, refusing AI reading turns every agent away, and making reads free stops charging agents. People always read free.
• A refusal is answered with a 403 that says why, on the article and through the REST API alike.
• A reader never waits on this: the rules are refreshed by a crawler request or the hourly check, and a person’s page view only reads the copy already stored.

0.5.5

• Your earnings figures are what the people on this site were paid. naulon’s fee is charged to the agent on top of an author’s share, and it was being added into the totals, so every figure read high by the amount of that fee and the fee wallet appeared in your list of authors.
• The fee is now stated on its own line beneath the totals, so you can see it rather than having a number quietly change.
• The toll test tells you what the agent pays, what your author gets, and which chain the money lands on by name. Each side of the split says whose it is.
• The table of who may do what describes each permission in a sentence. The capability name is still there if you need it for code.

0.5.4

• Your licence document is checked before it is published. A response that is not a licence, whether an interstitial from whatever sits in front of your site or an error page, is refused instead of being served as your terms, and the terms already published stay up.
• The Diagnostics check that names an unwritable plugin folder now asks WordPress’s own filesystem layer, so it answers correctly on hosts where WordPress does not write files directly.

0.5.3

• Exa’s crawler is now recognised. It identifies itself as a search engine, and for its consumer site it is one, but the same crawler feeds an API that hands an application the full text of your page from Exa’s cache, and a reader who gets your article that way never arrives at your site. Exa publishes one user agent for both, so there is no way to welcome the indexing and price the retrieval separately. It is charged by default.
• That default costs something, and you can change it. Charging Exa means dropping out of its index, consumer surface included. If reach matters more to you than revenue from that channel, set Exa to Allow on your Crawlers screen and it reads free, exactly like Google.
• Until now this plugin did not know the crawler at all, so it served Exa free whatever your account said.

0.5.2

• When an agent is asked to pay, the response now points at your licence. An agent that meets your toll already has that response in its hand. Before, it had to go back and read your robots.txt to find out what you allow and what it costs. Now it learns both in the same breath, which is one fewer reason for it to give up and read something else.
• Nothing is advertised that is not served: if your licence has not been published yet, the pointer stays off rather than sending a crawler to a page that is not there.

0.5.0

• Your site now publishes its licence. RSL, or Really Simple Licensing, is the open standard AI crawlers are starting to check before they read: a machine-readable statement of what a site permits and what it costs. Yours is served at /license.xml, pointed at from robots.txt and from the head of every page, and built from the price and scope already in your naulon account. Nothing to write and nothing to configure.
• Search indexers stay free in the licence, because they are free at the toll. What the document says and what the plugin does are the same thing, from the same settings.
• The licence is refreshed in the background, so a price change reaches crawlers within hours without you republishing anything. If naulon cannot be reached, the licence already published stays up rather than disappearing.

0.4.3

• Your dashboard now shows what the toll actually did. Every decision the plugin makes about a crawler, whether charged, served free, or re-read on a license it already paid for, is reported to your naulon account, so the Audit and Readiness screens stop telling a working site that nothing has been priced yet. Readers are still never reported: not sampled, not counted, not touched.
• Settlements are not reported from here. What you were paid is recorded from the payment itself, so nothing on your earnings screen can come from this plugin claiming it.
• A report that cannot be delivered, through a timeout or your site being offline for a moment, is kept and sent with the next one, rather than lost.

0.4.2

• Three more agent user-agents are now charged: Meta’s external fetcher, Amazon’s user-triggered fetcher and Mistral’s. They were reading your articles for free because the plugin only knew one of Meta’s five tokens. Search indexers are still never charged.

0.4.0

• Deleting the plugin no longer erases your data. WordPress removes a plugin’s data before it removes its files, so a delete that fails can still have wiped everything, and it did, on a real site. Your authors’ wallet addresses and your earnings record now survive a delete, and full removal is a box you tick in advance, on a screen that shows what it would destroy.
• Export your wallets and earnings to a file from naulon → Diagnostics, so nothing is one click from gone.
• Diagnostics now tells you if the plugin’s own folder is not writable by your web server, which is the reason an update reports that files could not be copied, or a delete reports that the plugin could not be fully removed. It names the folder and what to fix.

0.2.3

• A shared cache could hand a crawler the copy it made for a human, so a read that should have been paid was served free. The cache guard now varies on what the toll actually decides on.
• The toll test named two wrong causes and never the real one: a network edge in front of your site turning the crawler away before it ever reaches WordPress. It now names that, and what to change.
• The version in the plugin header, the readme and the code is checked against the last release, so a build can no longer ship under a version you already have.

0.2.0

• Refuse in-app enforcement on a domain the naulon fleet already tolls, so one read is never charged twice.
• Reconcile ownership on the hourly heartbeat: a proof withdrawn in the dashboard now stands the toll down here.
• A spent connect key reads as finished instead of failing, because the key drops its domain-management scope once the domain verifies.
• Clearer message when a key cannot claim domains, naming the key preset that can.

0.1.0

• First release: credits contract, author wallets, site ownership verification, roles and capabilities.
• Admin screens: setup with one-click verification and a real toll test, content policy, author payouts, earnings, diagnostics.
• Cache guard drop-in, plus a live check that tells you whether a crawler is actually being charged.
• Hourly heartbeat that keeps the connection alive and stands the toll down if DNS-based enforcement is already charging for the same domain.