Visibility – Lightweight SEO, AEO and GEO for Search and AI

For most sites, yes. It covers titles, meta descriptions, canonical URLs, Open Graph, schema, robots rules, redirects and the sitemap, and it imports your data from those plugins in one click.

It has no keyword analysis or content scoring. If you need those, keep your current plugin: Visibility is not meant to run alongside another SEO plugin.

1. Install and activate Visibility. Your old plugin can be active or already deactivated.
2. Open the Tools tab. The import card shows what it found, field by field.
3. Import, check a few entries and deactivate the old plugin.

What to expect:

• From: Yoast SEO, Rank Math, All in One SEO, SEOPress, The SEO Framework, Slim SEO and SureRank.
• What comes over: titles, meta descriptions, canonical URLs and robots rules of posts and terms, social titles, descriptions and images, schema types, verification codes and title templates.
• Site identity: entity type (with its organization or business type), name, logo, social profiles, contact, address and company details from SEOPress, Rank Math and Yoast SEO. Coordinates, price range and opening hours from Rank Math.
• More than one plugin: the one in use takes each value (active first, then installed, then the one set up last), and it decides the entity type.
• Redirects: from Rank Math, All in One SEO, SEOPress and the Redirection plugin.
• Templates: variables such as %%title%% are converted. One that needs a custom field or a date is left where it is.
• Safe to repeat: nothing is deleted from the old plugin, what you already set in Visibility is kept, and running it again never duplicates anything.

Install and activate Visibility, open the Tools tab and click Import now. The import copies your data and never deletes it, and what you already set in Visibility is kept. Then deactivate the old plugins.

It is free. There is no Pro version, no add-ons and no upsell banners. Everything on this page is in the plugin you download, including the redirects, the importers, the WooCommerce URL structures and the AI discovery files. No account and no license key.

It is built not to: no database tables of its own, no cron jobs and no external calls apart from a cached Gravatar check. Most of the work is done by WordPress core functions. The numbers of each release are in “Performance and security”.

No. There is no telemetry, no usage analytics and no account. The only outside request it can make is the cached Gravatar check described in “External services”, which sends a hash of the author email and never the address.

No. Settings live in one WordPress option, and per-post, per-term and per-user values in native WordPress meta. Deleting the plugin removes the option and keeps those values, so reinstalling does not lose your work.

Yes. The Overview tab switches Discover, Indexing, Sitemaps and Redirects on and off independently.

AEO (answer engine optimization) and GEO (generative engine optimization) mean being found, understood and cited by AI assistants such as ChatGPT, Perplexity, Claude and Google’s AI answers.

Nobody outside those companies knows exactly how they choose their sources, so Visibility works on what you can check yourself:

• Freshness: a filter and a Last updated column for content that has gone stale.
• Authorship: author details in Person schema and on the page.
• Questions and answers: FAQPage from the Q&A you have published.

No plugin can promise you a citation.

It gives AI search and assistants clean signals to read:

• Site Identity schema: who is behind the site.
• llms.txt and llms-full.txt: a Markdown index of your content.
• A Markdown version of every page at its .md URL.
• A robots.txt editor to allow or block AI crawlers by name.

The companion plugin VigIA shows which AI crawlers actually visit.

Yes, on WordPress 6.9 or later, which is where the Abilities API arrived. Visibility registers six abilities with it, and any client finds them: the WordPress MCP Adapter, Novamira and others.

• Read and write the SEO of a post, a page or a product.
• Read and write the SEO of a category, a tag or a term.
• List the entries with no Meta description or no Meta title.
• Read the SEO status of the site.

An agent can do no more than the user it is signed in as: each ability asks for the same permission as the editor. Visibility adds no server, endpoint or API key of its own.

To switch them off, untick the box in Visibility > Overview > AI agents.

Yes, both, generated on the fly from your content, so there is nothing to regenerate when you publish:

• /llms.txt: a Markdown index of your site, by content type and taxonomy, with WooCommerce-aware entries.
• /llms-full.txt: the long version.
• Every entry also answers at its own .md URL.

Both files are kept in wp-content/uploads/native-aeo-pack/, one per language, and rebuilt on the first request after a change. With the physical file option on, your server has to pass /llms.txt to WordPress when the file is missing.

Keep expectations realistic: llms.txt is a proposed convention, not a standard. Google has said it ignores it and no major AI vendor has confirmed support.

Yes, as a single JSON-LD graph, with per-post and per-term overrides:

• Article, with the type chosen per content type (BlogPosting, NewsArticle and others).
• BreadcrumbList.
• Person and ProfilePage for authors.
• FAQPage from Details and Accordion blocks.
• Site Identity (Organization, LocalBusiness or Person, and WebSite) on the home page.

It does not add Recipe, Product, Event or HowTo. Those need data that the plugins owning it already mark up, such as WooCommerce for Product.

Yes. In Discover > Site Identity pick Local business as the entity type:

• Business type: Restaurant, Dentist, Store and many more, or the generic LocalBusiness.
• Address and coordinates.
• Phone and email.
• Price range.
• Opening hours, with split hours.
• Cuisine and menu, for restaurants, cafés, bars and other food businesses.

Google lists the address as required for a local business. The settings screen warns you when it is missing.

It is one node on the home page, shared with the rest of the schema. It is not a store locator and it does not manage reviews.

Visibility describes one business. Give each other location a page of its own and add its node there with the native_aeo_pack_schema_graph filter:

add_filter( 'native_aeo_pack_schema_graph', function ( $graph, $context, $post ) {
    if ( 'article' === $context && $post && 'riverside' === $post->post_name ) {
        $graph[] = array(
            '@type'              => 'Restaurant',
            '@id'                => get_permalink( $post ) . '#business',
            'name'               => 'Casa Marta Riverside',
            'url'                => get_permalink( $post ),
            'address'            => array(
                '@type'           => 'PostalAddress',
                'streetAddress'   => 'Calle del Río, 3',
                'addressLocality' => 'Segovia',
                'addressCountry'  => 'ES',
            ),
            'parentOrganization' => array( '@id' => home_url( '/#identity' ) ),
        );
    }
    return $graph;
}, 10, 3 );

Google asks for each location to be a LocalBusiness of its own, with the address of its page in url.

In Discover > Site Identity:

• Name: the name of the site. Empty, it is the site title.
• Alternate names: an acronym or a shorter name, one per line.

Both go in the WebSite markup of the home page, which is where Google reads the site name from. That markup stays on the home page when you print the identity on another page.

The test only lists an organization that carries a contact detail. Add an email or a phone in Discover > Site Identity and it shows up.

Without one, the markup is still valid and Google still reads the name, the logo and the profiles. You can see it in the page source or in validator.schema.org.

Yes, at three levels:

• Everything: uncheck JSON-LD: Article and BreadcrumbList in the Discover tab.
• One content type: set its schema type to None.
• One entry: pick None in its Schema type field.

None adds no Article, BreadcrumbList or FAQPage to that content. Site Identity is a separate setting and keeps working. Developers can also use the native_aeo_pack_schema_graph filter.

Yes. The author archive of anyone with at least one published post gets a ProfilePage with their Person: job title, topics, credentials and profiles.

An author can also pick a page, such as About me, as their profile page, and that page becomes their ProfilePage. The Author details block or the [visibility_author_details] shortcode shows those fields on the page.

Yes. It customizes the sitemap WordPress already generates at /wp-sitemap.xml, so there is no second sitemap, no cache to clear and nothing to rebuild:

• Exclude post types, taxonomies, single posts and terms, and users by role.
• Noindexed content stays out automatically.
• Content whose Canonical URL points to another address stays out too. Google asks not to list one address in the sitemap and declare another as canonical.
• Add URLs of pages that are not WordPress content.
• Set lastmod dates and the number of URLs per sitemap.
• The sitemap URLs of your previous SEO plugin (sitemap.xml, post-sitemap.xml) are 301-redirected to the native one.

Because it answers with a 404 status while showing the right file, and Google discards any sitemap that does not answer 200. It happens on sites with no published blog posts: a shop, a portfolio, a site made of pages only.

Whether it affects you depends on your WordPress version: it was fixed in WordPress 7.1.1, from a report that came out of this plugin. On older versions, Visibility answers /wp-sitemap.xml with a 200 itself, with a switch for it in the Sitemaps tab. A sitemap with nothing in it still answers 404, as it should.

Something has switched off the sitemap WordPress generates, so the old address has nowhere to go. The usual causes:

• A snippet that sets wp_sitemaps_enabled to false, in functions.php or in a snippets plugin.
• Another SEO plugin serving a sitemap of its own.
• “Discourage search engines from indexing this site” ticked in Settings > Reading.

Visibility checks before redirecting and leaves the old address alone in that case.

Redirects are included and you need neither:

• 301, 302, 307 and 410, with no .htaccess editing and no custom table.
• CSV import and export.
• A suggested redirect when you delete a post.
• Take over or remove the redirects WordPress creates when a slug changes.
• A redirect whose target is redirected too is answered in one hop, and one that would close a loop is refused.

It does not log 404 errors, on purpose: that needs a table and a cleanup job. The companion plugin VigIA records the 404s that crawlers hit.

Yes. Its Meta Description Generation experiment saves the description it writes in the Meta description field of Visibility, so the page prints one meta description and not two. The descriptions it was printing on its own are copied to that field by the update that brings this in, where the field is empty.

Yes. The Visibility panel shows in the Block Editor sidebar, and as a meta box on the classic editing screen when a builder replaces the block editor.

With Elementor, the Markdown version of a page is built from the Elementor layout, so AI agents get the real text.

Yes:

• Products, categories, tags and attributes get titles, descriptions, social tags, schema and robots rules like any other content.
• The product gallery image is a fallback for og:image.
• llms.txt is WooCommerce-aware.
• Cleaner product and category URLs: see “Can I clean up my WooCommerce URLs?”.

It does not add Product schema: WooCommerce already does, with price, stock and reviews.

Yes, with all public post types and taxonomies. WooCommerce attributes need “Enable Archives?” switched on in their settings.

Yes. The Visibility panel has a live preview with three views: your Google result, your social card and an AI Overview citing your page. It updates as you type.

It shows what goes into your HTML, not a score: the final look depends on each platform.

Yes, with a template in the Discover tab. Combine these placeholders with your own words:

• %%title%%: the title of the page.
• %%sep%%: the separator you chose.
• %%sitename%% and %%tagline%%.
• %%page%%: the page number.
• %%taxonomy%%: the first category of the entry, or the product category of a product.

Example: %%title%% %%sep%% %%sitename%%.

• The front page, each content type and each taxonomy can have a template of their own.
• A Meta title typed on an entry or a term always wins.
• The meta description works the same way, with one more placeholder: %%excerpt%%.
• Custom fields and dates are not available, on purpose: they need a query on every page.
• A theme that prints its own title tag, without the WordPress title API, is not reached by the template.

Open the Visibility panel in the editor sidebar, or the Visibility meta box in the Classic Editor, and tick noindex.

You can also use Quick Edit or Bulk Actions in the post list, and in the category, tag and taxonomy lists.

• Block editor: turn on Keep out of search snippets in the Advanced panel of a Group block, or apply the format of the same name to selected words.
• Classic editor: use the button of the same name in the Visual toolbar, or nosnippet in the Text tab.

Visibility also leaves that text out of the descriptions and summaries it generates. The Markdown version and llms-full.txt keep the whole page.

The mark on a Group stops applying if you deactivate Visibility. The one on words stays in the content.

nosnippet tells Google to show no snippet and not to use the text in AI Overviews and AI Mode. Respect nosnippet (Discover tab) applies the same to the files Visibility generates. With it on, an entry set to nosnippet:

• is listed in llms.txt by title and link only,
• has no text in llms-full.txt,
• has no Markdown version,
• gets no FAQPage.

It is on in new installs. On an update it turns on only if that changes nothing your site already serves.

Yes. WordPress alone can answer 404 at /robots.txt on a subdirectory install or when its rewrite rules are stale. Visibility serves /robots.txt itself, so it always loads.

If another plugin writes a physical robots.txt, Visibility adds its rules to that file in a block of its own and leaves the rest untouched. /?robots=1 is redirected to /robots.txt.

Yes, and each site keeps its own settings:

• On a subdomain or mapped-domain network, each site serves its own /robots.txt and /llms.txt.
• On a subdirectory network there is one robots.txt at the domain root, owned by the main site.
• A physical file is written only where it belongs: a single site, or the main site of a subdirectory network.
• Is a physical robots.txt in the network root giving every site the same rules? On Apache and LiteSpeed, a button in the Discover tab routes /robots.txt through WordPress. Only a network administrator sees it.
• On nginx, the screen gives you the directive for whoever administers the server.

Activate Visibility per site or for the whole network.

Yes, with a folder or a domain per language. Visibility ships a wpml-config.xml, which Polylang reads too:

• The Meta title, Meta description and social title and description of each entry are translated with the entry.
• Robots, sitemap and llms.txt choices are the same in every language: set one on any translation and it applies to all of them. The Canonical URL is not shared, so each translation keeps its own address.
• With WPML, choices that already differ between languages stay as they are until you change one. Polylang makes them equal whenever an entry is saved.
• Title templates, the home page description and the Site Identity texts are translated in WPML > String Translation or Polylang > Languages > Translations. Save the Visibility settings once so they show up there.
• The site description in llms.txt and in the schema is the Site Identity description or, if that is empty, the WordPress tagline. Translate the one you use.
• Categories and tags: fill in their SEO fields on each translation.
• Each language gets its own og:locale, llms.txt, llms-full.txt and Markdown versions.
• A redirect applies to the exact path written in it. With a folder per language, add one per language (/old-page and /en/old-page).

• A redirect holds for every domain, because it matches a path. A relative target (/new-page/) goes to the domain asked for. It cannot send the same old path to a different page in each language.
• Leave Site Identity > URL empty: each domain then uses its own.
• Use the virtual llms.txt and robots.txt. A physical file in the site root is the same for every domain, so it lists the addresses of one of them.
• A physical .txt file with broken accents is sent by your server without a charset: set UTF-8 for .txt files in the server, or use the virtual files, which Visibility always sends as UTF-8.

• WPML with a folder per language: the sitemap of the default language lists the sub-sitemaps of every language, so the one in robots.txt covers them all.
• Polylang does the same itself. Exclusions, noindex and lastmod work in each language.
• One domain per language: each domain has its own sitemap and robots.txt. Submit each one in Search Console.
• There is no hreflang in the sitemap. WPML and Polylang add it to every page, and Google needs only one method.

In standard WordPress post meta (_native_aeo_pack_* keys). Term overrides go to term meta and author fields to user meta.

Yes. The Tools tab exports your settings and the per-entry and per-term values as one JSON file, and imports it on another site.

• Content is matched by slug, so it works between sites that are not clones.
• On import, choose between filling in only what is empty (the default) and replacing what is there.
• Nothing is created: values land on content that already exists.
• Redirects have their own CSV export in the Redirects tab. Author profile fields are not included.

Set a threshold in the Discover tab (90 days by default). Your content lists get a “Not updated in 90 days” filter with a count, and a sortable Last updated column. Set it to 0 to switch both off.

Why it matters: an Ahrefs analysis of 17 million citations found that the pages AI assistants cite were updated more recently, on average, than the ones ranking in Google’s organic results. Google’s AI Overviews were the exception.

Yes, in the Indexing tab:

• Remove the base of each taxonomy on its own: /product-category/, /product-tag/, /brand/.
• Choose a product URL structure WooCommerce does not allow: the product slug alone (/my-product/), the category and the slug (/t-shirts/my-product/) or the full category path.
• Old URLs are 301-redirected to the new ones, so a product is never reachable at two URLs.

Good to know:

• A page or a term that already uses a path keeps it.
• Removing a base adds rewrite rules for every term. The settings screen tells you how many before you save.
• If you use Premmerce Permalink Manager, import its structure from the Tools tab, check your URLs and deactivate it.

Yes, from content you already have. Switch on JSON-LD: FAQPage in the Discover tab and Visibility reads the Details and Accordion blocks of your entries: the question from the heading, the answer from the content.

It is off by default, because a Details block is not always a question. It helps AI assistants read your Q&A. It does not bring back Google’s FAQ rich results, retired in 2023.

Yes.

Read what Visibility resolved:

• Native_AEO_Pack_Frontend_Robots::is_noindex( $post_id ), is_nofollow( $post_id ), is_term_noindex( $term_id, $taxonomy ) and is_term_nofollow( $term_id, $taxonomy ).
• Native_AEO_Pack_Settings::get( $group, $key, $default ).
• Native_AEO_Pack_Settings::is_robots_ai_blocking() and get_robots_ai_agents().

Change what is printed, with eight filters:

• native_aeo_pack_schema_graph ($graph, $context, $object): the JSON-LD nodes before they are printed. $context is article, person, identity or website (the WebSite node on its own, on the home page of a site that prints its identity on another page). Return an empty array to print nothing.
• native_aeo_pack_identity_business_types ($types): the schema.org types offered for a local business identity.
• native_aeo_pack_identity_food_types ($types): the business types that take a cuisine and a menu.
• native_aeo_pack_document_title ($title): the Meta title before it reaches <title>. Return plain text, which is escaped afterwards, or an empty string to hand the title back to the theme.
• native_aeo_pack_canonical_url ($url, $object): the canonical of a listing. Return an empty string to print none. Single entries keep the core get_canonical_url filter.
• native_aeo_pack_meta_description ($description): the final description, used in the meta tag, Open Graph, the Twitter card and the schema.
• native_aeo_pack_og_image ($image, $post): the image chosen for a single entry.
• native_aeo_pack_schema_faq_pairs ($pairs, $post): the question and answer pairs before they become FAQPage. Return an empty array to skip it.

React to what an AI agent changed through the abilities, with two actions:

• native_aeo_pack_post_seo_updated ($post_id): after the SEO values of an entry changed.
• native_aeo_pack_term_seo_updated ($term_id, $taxonomy): after the SEO values of a term changed.