AgentedSEO featuresEvery SEO module you expect. Your agent works on the same settings.
Tired of a schema plugin, a redirect plugin and an llms.txt plugin on every site? AgentedSEO features cover titles, schema, sitemaps, redirects, llms.txt, IndexNow and site checks in one plugin, and your agent edits the same fields with a preview, your approval, a change log and undo.
- Every change loggedUndo one, or a whole batch
- Agent limited to SEORisky edits ask you first
- Works without an agentAll of it in your dashboard
- No AI credits to buyBring your own agent
Switching a module off keeps its settings and data.
One change is waiting in Approvals. The other two are in the change log, with Undo.
The same fields your agent edits. Whoever saves, it lands in one change log.
- Claude
- Cursor
- Codex
Hermes
One plugin per SEO job adds up quickly.
Plugin sprawl starts small. An SEO plugin for titles. A schema generator because the SEO plugin’s schema wasn’t enough. A redirect plugin. Then an llms.txt plugin, an IndexNow plugin and an AI crawler blocker. Then a bridge plugin, so your agent can reach any of them.
Every one of those is another settings screen and another update to test. One WordPress user put the wish plainly:
“I’d switch if another plugin could make the day-to-day SEO work noticeably simpler without adding a lot of extra stuff to the dashboard.”
AgentedSEO puts those jobs in one plugin.
As modules you switch on or off in Tools › Features. Your agent reaches all of them through its own login, with no bridge plugin.
SEO pluginSchema generatorRedirect pluginllms.txt pluginIndexNow pluginAI crawler blockerMCP bridge for the agent
AgentedSEO
- Core and safety
- Agent access
- Search appearance
- Schema
- Technical SEO
- Redirects and 404s
- Site check
- Editors and languages
- Importers
- 41
- active plugins
- 3
- SEO tools
- 2
- schema generators
- 6.2s
- to load on mobile
Every AgentedSEO feature works two ways, and both land in one change log.
Each setting is stored once. You change a meta description in the SEO drawer. Your agent changes it through MCP, the REST API or WP-CLI. An importer brings it over from your old plugin.
All three write the same field, pass the same checks and show up in the same change log, with the same Undo button.
That is why every module below has two columns: what you do in your dashboard, and what your agent can do through its own login.
“I built it this way for a reason. On classicmonks.com my agent did the SEO work, and I still couldn’t see at a glance what it had changed. Here, whoever makes the change, it lands in one log.”
JoyFrom the team behind Classic Monks · building WordPress sites since 2017See every change before it lands. Put back any change later.
Nothing on your site changes silently, whether you, your agent or an importer made the change. Risky changes wait for you, and anything that went through can be undone, one field or a whole batch. It’s the snapshot before and the diff after, built into the plugin.
Approvals, to read each waiting change with its old and new value. The change log, with Undo, Redo and CSV export.
Every write can run as a preview first. The agent gets the result of the checks, and when you reject a change, it gets your reason too.
What’s in it
- Preview on every write: the old value next to the new one, with the checks it passed, before anything is saved.
- Approvals with checks, a flag on any change that went stale since it was proposed, and emails when something is waiting. Approve all or Reject all, up to 100 at once, or one agent call’s whole batch together.
- Approval rules per area decide what waits for you. Some always ask, whatever you set: robots.txt edits, noindex, trash and deletes, and hiding a whole post type, taxonomy or the author archive from search.
- Change log with who made each change, when and how. Undo one change or a batch, redo after an undo, export to CSV, and set how long the log is kept. A settings reset can be undone too.
- Module switches in Tools › Features. Writes other plugins make to AgentedSEO fields are checked, logged and undoable as well.
- If a licence expires, your SEO output, data and undo keep working.
Your agent gets its own limited login, never your admin account.
Connect Claude, Codex, Cursor or Hermes and the agent signs in as its own WordPress user with the SEO Agent role. It can do the SEO work and write posts. It can’t reach users, plugins, themes, or header and footer code.
The Connections screen to create, revoke and self-test each connection. The Role and permissions screen. An agent activity log with CSV export.
72 actions, each available through the built-in MCP endpoint, the REST API, WP-CLI and the WordPress Abilities API.
What’s in it
Not in 1.0Agents editing Bricks or Elementor layouts. They can edit the SEO fields of those pages today.
- Connect with an application password per agent, or with OAuth 2.1 for the Claude.ai connector. An MCP self-test on the Connections screen shows the link works before you give the agent a job.
- Agents create, edit, publish, schedule and trash posts and upload media. Edits to live posts and publishing go through approval, undo works through WordPress revisions, and an edit waits instead of overwriting a post someone has open. You pick which post types agents may write.
- Sign-in guards: an address is locked out after ten failed sign-ins, agent users can’t sign in over XML-RPC, and revoking an agent hands the content it created to you.
- Prompt-injection guard: the server tells agents not to follow instructions they find inside your site’s content.
- For developers: an OpenAPI file, 19 filters and actions, template tags, a shortcode and blocks.
Set titles, snippets and share cards once per content type, not per post.
Write the pattern once for posts, pages, products or any content type, and every page prints its title, description, robots, canonical and share card from it. Override a single page in the SEO drawer when it needs it. No more typing title tags and meta descriptions one post at a time.
Meta templates with a live preview, Indexing rules, Social defaults and the Breadcrumbs builder.
The agent reads and writes the same titles, descriptions, robots and social fields, one page at a time or up to 100 in one bulk write.
Live preview of the result. Switch the content type above.
What’s in it
Limits in 1.0Max-snippet and max-image-preview are set per content type, not per page; terms get noindex and nofollow. There is no switch to hide the whole breadcrumb trail on one post.
- Meta templates and variables, the current year and date included, with a live preview of the result.
- Indexing rules per post type and taxonomy, with max-snippet and max-image-preview. A per-page index switch keeps one page indexed when its type is hidden (noindex still wins). Robots meta and canonicals, with self-referencing canonicals you can switch off.
- Social defaults and previews, up to 3 extra share images taken from the content, product price Open Graph tags on product pages, and oEmbed cards that use your social title and image.
- Breadcrumbs from a structure builder, placed with a block, a shortcode or a template tag.
- Image SEO on upload, empty alt text filled in when a page is shown, RSS settings, rules for links to other sites (nofollow, new tab, hosts to leave alone), and a switch to hide a post from site search and archives.
- Private content stays private: password-protected and WooCommerce Memberships content stays out of meta, social and schema.
A WordPress schema plugin built in, checked before anything saves.
No more pasting JSON-LD from a generator and hoping Google accepts it. Schema is checked against rich result rules when you or your agent save it, and invalid schema is refused. Every page prints one connected graph.
Site identity, defaults per content type, the Schema Rules screen, a form builder with code mode, and a FAQ block.
The agent writes schema through the same fields and gets the same check results, so invalid schema is refused before it saves.
One connected @graph on every page.
What’s in it
Not in 1.0Default schema for Elementor, EDD and LearnDash content types 1.2
- Validation for 22 schema types against rich result rules, with one connected @graph per page.
- Site identity: an Organization, or a Person linked to a WordPress user, a LocalBusiness telephone, and your logo with the site icon as fallback.
- Schema Rules match by content type, taxonomy, term, page template or front page. The first match wins, and you can preview a page a rule applies to. Defaults per post type and a page overrides table cover the rest.
- A form builder with code mode and saved templates, a custom site-wide JSON-LD box, and a FAQ block that also reads existing Yoast and Rank Math FAQ blocks.
- WooCommerce Product, ProductGroup and return policy, with brand, GTIN, sale end date and up to 5 reviews, and WooCommerce’s own duplicate JSON-LD removed.
- See what’s printed: the editor’s Schema tab shows what is printed on the page, with a Rich Results Test link.
Sitemaps, robots.txt, llms.txt, IndexNow and AI crawler rules in one place.
The crawl and index jobs people add separate plugins for are already in. Sitemaps, robots.txt, llms.txt, IndexNow pings and AI crawler rules sit on one set of screens. When your agent asks to change robots.txt, the change waits for you.
Sitemaps, the robots.txt editor with a URL tester, AI crawler toggles, llms.txt, IndexNow, and Verification and code.
The agent can propose robots.txt and llms.txt changes. Robots.txt edits always wait for your approval, and header and footer code stays off-limits to agents.
What’s in it
- XML sitemaps, plus video and news sitemaps, in every WPML or Polylang language. Password-protected posts, hidden products, cart, checkout and account pages are left out, and a Rebuild action redoes them.
- A robots.txt editor that never writes a file and keeps the lines other plugins add. Its URL tester shows whether Googlebot, Bingbot or GPTBot is blocked, and by which line, and a line that blocks your sitemap gets flagged.
- AI crawler toggles that split search crawlers (OAI-SearchBot, PerplexityBot, Claude-SearchBot) from training crawlers (GPTBot, Google-Extended, ClaudeBot, CCBot), so you can stay in AI answers and opt out of training. They write robots.txt rules, which crawlers that respect robots.txt follow.
- An llms.txt file you or your agent can edit. It isn’t served while search engines are discouraged, and it’s sent with a noindex header.
- IndexNow: the key is served for you, changed URLs are sent on publish and update, removed pages are reported, and you get a warning if another plugin also pings.
- Verification codes for Google, Bing, Pinterest and Yandex sit next to head and body code snippets.
Redirects and a 404 monitor, with no separate WordPress redirect plugin.
- A redirect manager that refuses loops and chains, takes regex (unsafe patterns are refused), and imports and exports CSV. Rules from the Redirection plugin import too.
- An automatic 301 when a slug changes, child pages included. Rules that pointed at the old URL move to the new one, and the rule is removed if the post moves back.
- A 404 monitor with a “New this week” count and a count of 404s linked from your own pages.
Not in 1.0Suggested redirect targets for each 404 1.1
A site check that names each problem and fixes the setting behind it.
You get a short list of real problems with a Fix button, not a green signal to chase. Each issue has a stable code, so you, your agent and your reports mean the same thing by it.
The Issues screen, the Overview with a site score, the setup checklist, Broken links, Image alt text, and bulk titles and descriptions.
The agent reads the same issue codes and can run a fix as a preview first. The fix then follows your approval rules like any other change.
Every fix is logged, so it undoes like any other change.
What’s in it
Not in 1.0The internal link graph and link suggestions 1.2. Search Console checks are coming soon.
- An Issues screen with every 1.0 check code, filters and CSV export. One-click fixes for site-level issues: a missing sitemap, robots blocking, noindex conflicts, share-card problems, a missing site entity, HTTPS and canonical issues. Redirect chains and noindex pages in the sitemap are checked too.
- Broken links, internal and external, with recheck, unlink and a daily email of new broken links. 401, 403 and 429 responses don’t count as broken.
- An Image alt text screen and a bulk titles and descriptions screen, for fixing many pages in one sitting. Images that appear only inside post content are shown read-only there.
- Content checks read the rendered HTML of Bricks, Elementor, Breakdance and Oxygen pages, plus ACF fields. A keyword plan warns you when two pages target the same keyword.
- Where you already look: a WordPress Dashboard widget, an admin bar count of open issues, a monthly SEO report email, and a posts list you can sort by open issues.
The SEO drawer opens in your editor, your page builder and every language.
No switching screens to do SEO, whatever you build with. The same drawer opens in the block editor, the classic editor and the posts list, and inside Bricks, Elementor, Breakdance and Oxygen, in every WPML or Polylang language.
Opens from a floating button where a builder has no toolbar spot.
SEO, Social, Schema and Advanced tabs on every post, plus term fields, a posts list column and Quick Edit.
With “I work with an AI agent” switched on, an Agent / You switch above the drawer tabs shows what the agent changed or proposed for this post, with a link to Approvals.
With WPML or Polylang, SEO fields carry into each new translation and every language gets its own sitemap.
What’s in it
Not in 1.0A Bricks breadcrumbs element 1.1. Agent changes waiting for you show in Approvals, not yet in the document summary or the pre-publish panel.
- SEO, Social, Schema and Advanced tabs, plus a document summary and a pre-publish panel in the block editor.
- In a page builder with no toolbar spot, the drawer opens from a floating button. The posts list and builders warn you before unsaved SEO changes are dropped.
- Term edit screens get SEO fields and the archive’s own schema.
- WPML and Polylang: sitemaps and alternate locales in every language, SEO fields carried into new translations, and site-wide text translated per language.
- Checks run in each post’s own language. Readability runs on English posts only.
Bring your SEO from 10 plugins, previewed first and undoable.
Switching shouldn’t mean redoing years of titles, descriptions and redirects, or finding out later that some pages lost their keyword settings. A dry run shows how many values each field would get, plus conflicts and skips, before anything is written. The whole import undoes in one click.
The importer, with a dry run, a choice to keep or overwrite values a page already has, Undo, and an offer to deactivate the old plugin when you’re done.
Once the import is in, your agent works on the imported titles, descriptions and redirects like any other field.
What’s in it
- Per-page SEO from all 10: titles, descriptions, keywords, canonicals, robots, social, primary term and redirects.
- Site settings: meta templates and separator, sitemap on or off, social defaults and site identity from all 10. Verification codes, robots.txt, breadcrumb settings and indexing rules come from the plugins that store them.
- Redirect CSV exports from Yoast, Rank Math, Slim SEO and SEOPress, up to 5,000 rows, undoable.
- Keep or overwrite values a page already has. Undo leaves alone any value you edited after the import, and your old plugin’s data is never touched.
- What doesn’t come across: SEO and readability scores, cornerstone or pillar flags, noimageindex and a few rare robots tags, and per-post schema, since the post type defaults cover it. Settings with no match show up as notes in the dry run.
- Settings export and import between your own sites.
Install, import, pick your modules, and add an agent if you want one.
- 01
Install and run the setup wizard
Add AgentedSEO to your site and let the wizard walk you through the basics.
- 02
Bring your old SEO across
Pick your current plugin, preview what moves, and import. One click undoes it.
- 03
Switch on the modules you want
Tools › Features turns each module on or off, so you leave the rest out of your way.
- 04
Add an agent, if you want one
Paste one URL into Claude, Codex, Cursor or Hermes. The agent works through its own role, and you approve from Approvals.
What isn’t in 1.0 yet, so you buy on what’s built.
Release labels, no dates. Lifetime licences get every one of these as an update.
- Search ConsoleClicks, positions, opportunities and URL inspection inside WordPress. Switched on once the connection service is live.
- Suggested redirect target for each 404
- Keyword position history charts
- A Bricks breadcrumbs element
- Internal link graph, orphan pages and link suggestions
- Planning fields for the keyword plan: intent, priority, status, owner
- A Bing Webmaster Tools connection
- A per-area role matrix for your team
- Default schema for Elementor, EDD and LearnDash content
- GA4 numbers per page
- Agents editing Bricks and Elementor layouts
Questions people ask before they switch SEO plugins.
Still unsure? Try it in a sandbox.
For the jobs a schema plugin usually does, yes. It validates 22 types against rich result rules before saving, prints one connected graph per page, and covers site identity, per-type defaults, Schema Rules, a form builder with code mode, a FAQ block and WooCommerce products. Default mappings for Elementor, EDD and LearnDash content types come in 1.2.
Not for redirects and 404 tracking. AgentedSEO has a redirect manager with regex and CSV import and export, an automatic 301 when a slug changes, and a 404 monitor. Rules from the Redirection plugin import. Suggested redirect targets for each 404 come in 1.1.
Yes. You or your agent edit llms.txt in the plugin. It isn’t served while search engines are discouraged, and it’s sent with a noindex header so the file itself stays out of search results.
Yes. It serves the key for you, sends changed URLs when you publish or update, reports removed pages, and warns you if another plugin is also pinging IndexNow.
Yes. The AI crawler toggles treat search crawlers such as OAI-SearchBot and PerplexityBot separately from training crawlers such as GPTBot and CCBot. They write robots.txt rules, so crawlers that respect robots.txt follow them. It is not a firewall: a crawler that ignores robots.txt isn’t stopped.
No. Every module works from the dashboard. An agent is how you hand off the repetitive parts, like titles and meta descriptions across 300 posts, when you’re ready.
No AI credits. The optional AI helpers are off by default and run on your own key: OpenAI, Anthropic, Gemini, OpenRouter or any OpenAI-compatible server, local ones included. Site-wide limits default to 100 a day and 10 a minute.
They can edit the SEO fields of any page, Elementor and Bricks pages included. The layout itself stays off-limits to agents for now. More on the home page FAQ
Coming soon. The Search Console screens are built and switch on once the connection service is live. Lifetime buyers get them as part of their updates.
Early Adopter Lifetime doesn’t expire, and its updates and support last for life. On a yearly licence, if you stop renewing, your SEO output, data and undo keep working. Only updates and support stop.
Run your SEO from one plugin. Let your agent do the repetitive parts.
Early Adopter Lifetime. One payment, lifetime updates. Closes 1 December, 23:59 PT.
15-day refund (80% back)Questions? [email protected]
Built by Joy, from the team behind Classic Monks. 64% of the Classic Monks community poll (34 votes) asked for SEO as its own plugin.