Linux 软件免费装

在这里免费安装WordPress以及其他软件

Banner图

Agentimus – AI SEO, llms.txt & MCP for AI Agents

开发者 heera
更新时间 2026年8月26日 00:27
PHP版本: 7.4 及以上
WordPress版本: 7.1
版权: GPLv2 or later
版权网址: 版权信息

标签

ai-seo llms-txt mcp ai-agents agent-readiness

下载

1.34.0 1.37.0 1.5.0 1.9.0 1.22.1 1.39.0 1.15.0 1.16.0 1.3.0 1.4.0 1.2.0 1.4.1 1.4.2 1.4.3 1.0.0 1.12.4 1.14.0 1.14.1 1.18.0 1.19.0 1.6.0 1.10.0 1.13.0 1.20.0 1.20.1 1.21.0 1.21.1 1.21.2 1.23.0 1.17.0 1.25.0 1.26.0 1.27.1 1.28.0 1.29.0 1.30.0 1.10.1 1.32.0 1.22.0 1.24.0 1.27.0 1.35.0 1.36.0 1.22.2 1.38.0 1.40.0 1.40.1 1.41.0 1.42.0 1.1.0 1.43.0 1.33.0 1.44.0 1.8.0 1.31.0 1.45.0 1.46.0 1.7.0

详情介绍:

Agentimus is an all-in-one AI SEO plugin for the age of AI agents — AEO (Answer Engine Optimization) and GEO (Generative Engine Optimization) in one place. It does two things. It makes your site legible and citable. It helps AI assistants like ChatGPT, Claude and Perplexity find your site, read it correctly and cite it in your own words — and shows you which AI bots are visiting. You don't need to understand AI or web standards: a setup wizard walks you through it in about a minute, then it runs on its own. And it lets the AI tools you already use operate your site. Turn on the built-in Model Context Protocol (MCP) server and Claude Code, Claude Desktop, Cursor or Codex can read your reports and — behind two more opt-in switches — draft, edit and publish posts. All three are off by default. Prefer wp-admin? A built-in writing assistant drafts and revises posts there. By default it makes no outbound requests, collects no analytics and logs no IP addresses — everything runs on your own site. Three optional features change that only when you switch them on: Citation checks, Verify bot identities and Store IP addresses, each disclosed in full under External services. 📖 Full documentation — a plain-English manual and developer reference, step-by-step guides for every feature: https://heera.github.io/agentimus/ One screen that says what needs you With an SEO plugin — or instead of one Operate your site from your AI agent (MCP) — opt-in Write with AI in wp-admin — the built-in assistant (opt-in) Tell your team — and announce what you publish Control — who may use your content Visibility — who is reading you Classic search, measured — Bing & Google (opt-in) Content — clean, machine-readable output Identity & contact Readiness report Machine discovery (forward-looking) Everything above is read by search engines and AI tools today. This part is forward-looking — the conventions the agent ecosystem is converging on, putting identity, capabilities and APIs in one predictable place: Why it's useful Most tools cover one slice — an llms.txt file, a bot blocker, or structured data. Agentimus brings the whole job together in one lightweight package — and tells you what's still missing.

安装:

  1. Upload the agentimus folder to /wp-content/plugins/, or install via Plugins → Add New.
  2. Activate the plugin.
  3. A setup wizard opens automatically on your first visit to the admin and walks you through your identity and content choices in about a minute. After that everything runs on its own — open Agentimus any time to review the readiness report or adjust settings.

屏幕截图:

  • Visibility → In the index — whether the engines' indexes actually hold your pages, Bing's and Google's on one screen. Bing is the index ChatGPT search and Copilot read today: how much of your site it holds, how cleanly its crawler gets in, and a live question you can ask about any single page. Google's is walked in rotation and re-checked daily — healthy pages stay a count, while anything that needs a look earns a row in Google's own words, each with a link straight to that URL in Search Console.
  • Visibility → Search — what people searched, how often you appeared, how often those results were clicked, and the pages that earned them; every number the engine's own. Connect both Google and Bing and a switch appears, because the two count different searchers and are never merged. Directly beneath it, Search Opportunities turns those same numbers into a worklist: pages sitting just off page one, and pages on page one being scrolled past, each wired to the exact field that fixes it. It also names the searches several of your pages are splitting between them — the clicks divide, so each ranks lower than one page would — with the page that earns the click stated, and one decision on every other.
  • Readiness report — a plain-English pass/warn/fail checklist of what's enabled and what's still missing, grouped by rung, each row carrying its own fix in the site's own words. Buttons at the top preview your site the way an agent sees it, re-run every check live, and scan for files that shouldn't be public. Beneath the checklist your content is summarised rather than listed: how many pieces are graded, how many carry something worth fixing, and which issue is the most common — with a way in page by page, a way in by issue, and the set-aside list of everything you have deliberately left out of the score.
  • Crawler policy & bot identity — declare your content-usage signals, block scanners and scrapers by name, turn away spoofed traffic, and verify bots against what each operator publishes: reverse DNS for the search engines, published IP ranges for GPTBot and PerplexityBot. The verified-bots registry is yours to edit.
  • Request Log — every request an agent made, in one filterable table: narrow by client, endpoint, network, verification verdict, User-Agent or date to see exactly what a single bot fetched. A Status column carries one honest mark per request — verified, signed, spoofed, forged, refused or unchecked. Your own logged-in visits are never recorded.
  • MCP server — one switch runs a Model Context Protocol server on your own site. Your server address leads: give it to an assistant and it asks you for approval — no keys to paste — and every approved assistant is listed with its scope, its last call and its own Disconnect. Above it, the trust ladder: a second switch lets agents write, a third decides whether they may publish or only leave drafts.
  • Data sources — the outside services Agentimus reads from, all optional, each with one key held on your own server and no third-party proxy: Cloudflare for what the edge answered or blocked before your server saw it (and, with an optional extra permission, clearing its cache when you publish), Google Search Console for classic search and index coverage — with Analytics alongside it, on that same key, if you want visitor numbers too — and Bing Webmaster Tools, which can also announce every change you publish through IndexNow. Numbers land in your own database, so your history outlives each service's own reporting window.
  • Visibility → Citations — an opt-in, bring-your-own-key scoreboard showing whether ChatGPT, Perplexity, Gemini and Claude mention and link each brand, product or person you track: seen-in-answers and linked-your-site rates, rank against each item's own rivals, and question-by-question results with the sources each engine cited.
  • Discovery — everything your site offers AI assistants, in one place and in plain words: the things assistants can read, the things they can do, and what each of those rows actually allows. Beneath them the MCP connection, and every well-known document your site serves with its live status and one line saying what it is for — so none of it is a filename you have to go and look up. Any registration problem is listed with a plain-English fix.
  • The writing assistant — a spark button on every Agentimus screen opens the drawer: choose what you are writing (a post, a page, or one of your own content types — a page gets no invented sections and no tags), describe it in your own words, then draft it straight away or shape the outline first. Preview the fully dressed draft, create it as a draft and land straight in the editor. Nothing is saved until you say so, and it never publishes.
  • In the post editor — the "Agentimus" box, Readability tab: a per-page pass/warn check of what makes the page hard to read, section and quote — for AI assistants, for search engines and for people alike — enough substance, an opening summary, specifics and cited sources, section headings and their order, quotable passages, reading ease, prose vs links, image alt text, video and audio, the featured image and whether it is described, and freshness. The rows an AI can draft a fix for — a thin page, a missing opening summary, an over-long block — carry a "Fix with AI" button once you have set a provider up in WordPress. Two more tabs sit alongside it: JSON-LD and Share.
  • Findings — one front door for everything open across your site, ranked by what each one costs: pages losing a click they already earn, and anything the setup checks caught. Every row says what to do and carries the button that lands on the fix, and anything filed under “when you have time” can be put away with a count of what you put away. Beneath it, your content one row at a time — what each page is found for, whether it answers that, and anything else it needs. It covers every kind of content you publish, and the line above the list says which kinds are being checked.
  • Integrations → Services — where Agentimus sends its reports. A new finding, a caught impostor, the weekly digest and three more moments can go to Telegram, Slack, Discord, a Google Sheet, or any address of your own as signed JSON. Each service is one you own — your bot, your channel, your sheet — and every connection is tested before it is saved, so a card never claims to be connected when nothing would arrive. All of it is off until you switch it on.
  • Integrations → Plugins — what your other plugins tell AI assistants about your site, on your behalf. A WooCommerce store says it sells products; FluentCart says it has store and checkout pages; FluentCommunity says it holds community spaces. Plugins that keep everything behind a login are listed too, saying exactly that — a plugin with nothing public has nothing to pass on, and naming it is more use than leaving it out. Any plugin can describe itself the same way through one hook, with no dependency on Agentimus.
  • Report — what AI did on your site between any two dates. How many AI crawlers read you and which ones, how many people arrived from an AI answer and from where, what assistants did here, where you stand in search, your score, and the one thing worth doing. Today, yesterday, seven days, thirty, or two dates from a calendar. Every block says how fresh it can be: your own log answers any window to the minute, while Google and Bing publish days behind and name the newest day they have rather than printing a zero that would read as "nobody searched". Your dashboard opens with the same reading for today.

升级注意事项:

1.46.0 A connected assistant can now maintain a site rather than only rewrite it: read a page, change one passage, and describe pictures that have none. Two alt-text faults are fixed — the tool that describes a picture used to refuse the exact pages the checks flagged, and an empty description on a captioned picture now counts as a gap. Your site re-reads its pages once after updating, so a finding that was true all along may appear. No breaking changes. 1.45.0 Two readings that could be quietly wrong are fixed. A page could be listed as not answering a search it is no longer found for; those pages are read again now when their search changes. And a day Bing reported nothing about was drawn as a fall to zero pages in the index — those days now show as a gap that says so, and the index figure above the chart can no longer read zero because Bing stayed quiet. No breaking changes. 1.44.0 A new Report screen under More says what AI did on your site between any two dates, and the dashboard now opens with today. Citation sources name the site that was actually cited instead of Google's redirect, so "linked your site" is a real reading at last — it may go up on your first run without anything changing on your site. You can also click a verdict and read the whole answer. A heading that was invisible on dark admin schemes is readable again. No breaking changes. 1.43.0 The Google index check no longer restarts itself after finishing, and Cancel really stops it now. Publishing also clears your listing pages from Cloudflare — the page your articles are listed on, category and tag pages, your feed and your sitemap. Links on the dashboard no longer look identical to status text. A connected assistant can see the clients waiting on a verdict and answer them. 1.42.0 Your dashboard no longer shows Agentimus's internal names for the files AI assistants fetch, and the sidebar's two lists say what they hold. Plugin cards and open Discovery addresses are now links you can follow. A connected assistant can read your announcement ledger and see what you are connected to — without ever being handed an address or a key. Also fixes a case where another plugin's runtime change could be saved into your own settings. No breaking changes. 1.41.0 Agentimus keeps its own colour on WordPress 7.1's default admin scheme — 1.40.0 turned its charts, links and numbers black there. A connected AI assistant can now set a page aside as one you never meant to be quoted, one page at a time, and can filter your request log by several crawlers at once. Titles you type yourself are no longer recapitalised. No breaking changes. 1.40.1 Fixed a bug regarding the request log as a connected AI assistant reads it: 1.40.0's new sorting added two fields the assistant-facing tool had not been told about, so it refused its own reply and an assistant got an error instead of your log. Nothing you see in the browser was affected. No breaking changes. 1.40.0 Fixed bugs regarding the impostor count and the dashboard's drill-downs: your site could report "0 caught faking an identity" while it was catching and refusing them, and clicking a number opened an unfiltered list. Two Light-scheme faults are fixed too — an accent colour nothing could read, and a row that vanished when you hovered it. The Request Log's filters now take several values at once and every column sorts. No breaking changes. 1.39.0 Your content now gets read even if your host never runs WordPress's scheduled jobs — many never do, and nothing used to say so. Working through one issue now opens every page that check flags rather than six of them, and a page you fix says so instead of vanishing from the list. Several counts that disagreed from one screen to the next read the same measure now, so a few of your numbers will move once. No breaking changes. 1.38.0 Your content is read again after this update — a verdict now records which checks made it, so expect your counts to move once: pages that read as clear were answered by checks the plugin no longer runs. Your lists stay on screen, marked, while that happens, and a page you half-fix no longer vanishes from Optimize Your Content. The featured-image check now reads a page you actually serve. Connected assistants gain three tools: the pages worth fixing, your real categories and tags, and the two fields behind a search result. No breaking changes. 1.37.0 Your Optimized score now reads your whole site, not its 25 most recent edits — on most sites the number will move, and more pages will be listed as worth fixing. Two new checks can add warnings too: alt text that is only a file name, and a featured image with no description. Nothing broke; nothing was looking at any of it before. Also: reports to Telegram, Slack, Discord, Sheets or your own address; announcing to X and LinkedIn; your plugins described to AI assistants; the editor's AI Readability panel is now Readability. No breaking changes. 1.36.0 Findings — one front door for everything open across your site, ranked by what each thing costs you, every row carrying the button that lands on the fix. Light and dark, hand-tuned and keyed to your admin colour scheme, following your device unless you say otherwise. Search Opportunities moves under the search report it reads from, on Visibility → Search, and Readiness keeps a pointer to it. Bing gains a daily traffic series and a live per-page question; Analytics can ride the Search Console key you already connected. Every data screen re-reads when you come back to it. No breaking changes. 1.35.0 Ask AI in the post editor — one block, a selection, or the whole post as a list of proposed edits you accept one at a time. The writing assistant now writes pages and custom types as pages, not as articles. In Google's Index gains a live Re-check for a single page and a Cancel for a running check. Your llms.txt, discovery and agent-card documents are offered to connected assistants as readable resources, and an agent can search your media library. One change to know about: editing an existing post inside the assistant's drawer is gone — revision happens in the editor now. No content is affected. 1.34.0 IndexNow announcements the moment you publish (off by default), Cloudflare cache purge on publish and on demand (one optional token permission), and an In Google's Index card that reports the whole site rather than a watchlist — pages Google left out are re-checked daily until they get in, and it says so when they do. Plus Bing's own record of your sitemap, and every search card can now be set aside, even the homepage. No breaking changes. 1.33.0 Classic search, measured: Search Performance and Opportunities for Bing and Google, an In Google's Index card with whole-site coverage and page lookup, a Cloudflare edge view, and a setup wizard. The sitemap returns to WordPress's standard /wp-sitemap.xml — old addresses redirect, so registrations heal on their own. No breaking changes. 1.32.0 Connect an AI assistant by approving it, not by pasting keys: give it your MCP address and it asks you for permission on your own site, where you choose read-only or read-and-write. Each assistant gets its own key and its own Disconnect. Plus a shared token for clients that can't ask, honest reporting when another plugin serves your llms.txt, and a robots.txt change watch. No breaking changes — existing keys keep working. 1.31.0 Share drafts written from the post (six networks, editor-only, nothing auto-posted) and an "Ask AI about this post" row whose assistant visits show in your own request log — both with an off switch. Plus consistent loading screens and an MCP request-log fix. No breaking changes. 1.30.0 Agents that cryptographically sign their requests (Google's agent, OpenAI) are now verified on your own server — forgeries flagged, unsigned crawlers never penalised. Plus internal-link suggestions in the editor, refused impostors on the record, and a clearer Request Log. No breaking changes. 1.29.0 Agentimus now covers the search basics — titles, share cards, canonicals, sitemap — when no SEO plugin is active, and keeps deferring when one is. Plus: pick the weekly email’s day and time. 1.28.0 A short weekly email about what AI did on your site — quiet weeks send nothing, one-click stop inside every email — and the review queue’s count now shows on the admin menu from every screen. 1.27.1 Small ordering fix on the new review-ask endpoint (deny before validate). No security impact, no breaking changes. 1.27.0 A quiet, one-time review ask that never nags (and retires the footer's rating line once answered), a first-run finish that shows what just went live, and standard auth vocabulary on the MCP server card. No breaking changes. 1.26.0 Fairer AI-readability grading — no more penalty for your topic’s vocabulary or your code samples — and AI agents now draft to your readability rules, graded on every write. 1.25.0 Write with AI without leaving wp-admin: a built-in assistant drafts new posts (outline first, preview always, never publishes), revises existing ones without touching their status, and generates images right in the editor. Plus plain-language AI errors and a self-updating Cited rung. No breaking changes. 1.24.0 Bot verification grows: bots are also checked against their operator's published IP ranges (GPTBot, PerplexityBot, …), the verified list is now yours to edit, and with blocking on a proven impostor is refused — not just flagged. Plus a clearer review queue and a phone-friendly admin. No breaking changes. 1.23.0 Agent-access alerts now fire on every visit, Markdown output is ready for WordPress 7.1's new blocks, and your Content Guidelines steer every AI draft — in the editor and over MCP. No breaking changes. 1.22.2 Polish and fixes: clicking a Discovery tile now lands the section in view, the MCP settings card no longer logs a 401 in the browser console, the Discovery layout is tidier, and AI Visibility explains why it uses its own keys. No breaking changes. 1.22.1 New opt-in write tier: connected AI tools can draft, edit and fully dress posts — categories, tags, featured image — and, behind a separate third switch, publish. Off by default, permission-checked per key, every write audited. Plus: the MCP card now connects your AI tool for you, and the score card counts what's left to fix. 1.22.0 One switch now runs an MCP server on your own site — AI tools like Claude Code can use Agentimus's read-only tools, authenticated, permission-checked and fully audited. Off by default. Also: Agent access rows now name the user and key behind every event, and the dashboard's 7/30-day counters count whole calendar days, so they no longer appear to lose hits.

常见问题:

Where is the documentation?

The full documentation — a plain-English user manual and a developer reference — is at https://heera.github.io/agentimus/. It has step-by-step guides for every feature, plus the hooks, filters and endpoints for developers.

Do I need to be technical to use this?

No. A setup wizard opens automatically the first time you visit the admin and walks you through everything in about a minute — you write a sentence about who you are and tick what AI assistants may read. Everything else runs on its own, and you can change any of it later.

What does Agentimus change on my site? Will my visitors notice?

Nothing your visitors see changes — there's no new front-end script, style or layout. Behind the scenes it publishes machine-readable files and signals (like llms.txt and a discovery document) that only AI assistants and crawlers read. It also stands down automatically next to SEO plugins, so it won't duplicate or fight your existing setup.

What's the quickest way to set this up for my site?

Activate Agentimus and run the one-minute setup wizard — that covers most sites. Then, depending on what you do:

  • Consultant, freelancer or personal brand: fill in your Identity — your name, a one-sentence bio, your expertise topics, and links to your other profiles. That's the highest-signal information an AI assistant uses to describe and cite you correctly.
  • Business or agency: set the entity type to Organization, list the services you offer, and add a contact email so an agent can point enquiries the right way.
  • Blog or publisher: the defaults are already right — your posts and pages flow into llms.txt automatically. Just add a profile sentence so an assistant knows whose site it is.
Whatever your case, the Readiness report always tells you the single next thing worth improving.

Does Agentimus make external requests or send my data anywhere?

By default, no — Agentimus makes no outbound HTTP requests out of the box, sends nothing to any external service, collects no analytics or telemetry, and stores the agent-activity log in your own database with no IP addresses. (One opt-in setting, Store IP addresses for flagged clients, can store IPs locally for flagged crawlers only — off by default; see External services.) Two opt-in features go outbound. Verify bot identities makes DNS lookups and downloads, once a day, the crawler-IP lists that bot operators publish (Google, OpenAI, Perplexity, …) so impostors can be caught; when a request arrives carrying a Web Bot Auth signature, it also fetches that operator's public key file so the signature can be checked on your own server. Only those public files are fetched, and nothing about your site is sent. The other is the optional citation checks feature: if you enable it and add your own API key, Agentimus queries the AI provider(s) you chose (OpenAI, Perplexity, Gemini and/or Anthropic) to check whether they mention and cite you — only for the engines you turn on, and only when a check runs (on demand or on your schedule). Your keys stay on your server and nothing else is sent anywhere. See External services for the full disclosure. The discovery document includes a $schema value that identifies the document format (the same way a schema.org URL identifies a vocabulary); it is a label in the output, never fetched. The one place a request is made is the optional "Verify live" self-check on the readiness report — and that runs in your browser, fetching your own public URLs only when you click it; the server itself still makes no request.

Do I need an SEO plugin alongside Agentimus?

No. With no SEO plugin installed, Agentimus covers the search essentials itself: per-page SEO titles (an "SEO title" field in the editor), Open Graph/X share cards, canonical links, meta descriptions, and an XML sitemap with last-changed dates. Each has its own switch under Settings → Discovery → Search basics. If you install an SEO plugin later, Agentimus notices on the next page load and steps aside on everything the plugin owns — your per-page values are kept, and everything returns if that plugin leaves.

Does this conflict with my SEO plugin?

No. When Yoast, Rank Math, SEOPress, AIOSEO or The SEO Framework is active, Agentimus stands down on every overlapping surface — JSON-LD, SEO titles, share cards, canonicals, meta descriptions and the sitemap — automatically, so nothing is ever emitted twice. A card on the dashboard names the division of labour. The AI-facing endpoints (llms.txt, markdown twins, the discovery documents) don't overlap with SEO plugins and keep working the same either way.

My robots.txt rules aren't showing.

If a static robots.txt file exists at your site root, or your CDN serves its own, it overrides WordPress's virtual robots.txt. The readiness report flags this. Remove the static file to let Agentimus manage the rules.

An SEO audit tool says my robots.txt is invalid.

Some audit tools flag the Content-Signal: line Agentimus adds. The line is valid: the robots.txt standard (RFC 9309) tells parsers to ignore any field they don't recognise, so every crawler still reads the rest of your file exactly as written. The tool is behind the standard, not your site. Nothing to fix.

I turned something on but nothing seems to happen — is it broken?

Almost always it's working — here's how to confirm. The generated AI files are cached for up to an hour, so a change may not show instantly: open the file directly (for example yoursite.com/llms.txt) and refresh. The Readiness report's Verify live button fetches your real URLs from your browser and shows exactly what an agent receives — including anything your CDN is caching. If a file still isn't appearing, check that a static file or your CDN isn't overriding it (the report flags a static robots.txt, for instance).

How do I tell AI not to train on my content?

Set Allow AI training to off under Settings → Crawler policy. That one switch publishes your choice in three places at once, so a crawler that ignores one still sees the others:

  1. robots.txt — a Content-Signal: … ai-train=no line (advisory).
  2. A response header on your pages — tdm-reservation: 1 (the W3C TDM Reservation Protocol), which reaches bots that never read robots.txt.
  3. An opt-out file at /.well-known/tdmrep.json — the recognized, machine-readable reservation, relevant under EU text-and-data-mining rules.
The header and file are on by default and can be toggled per channel under "Published beyond robots.txt". You can optionally also send the non-standard X-Robots-Tag: noai, noimageai (off by default, honored by some platforms) and link an AI-usage policy URL. Important — these are signals, not a wall. robots.txt, the header and tdmrep.json are standardized requests that compliant crawlers honor; they do not forcibly stop a bot. To actually refuse a crawler with a 403, add it to the crawler list or use scanner blocking (Crawler policy → Block specific crawlers / Block scanners), which Agentimus enforces at its generated endpoints.

Can I block only specific AI bots?

Yes — list them under Block specific crawlers. That writes a per-name Disallow: / to robots.txt for each. The /.well-known/tdmrep.json opt-out file and the tdm-reservation header are site-wide — the standard has no per-bot dial — so per-bot blocking lives in robots.txt (and in scanner blocking for a hard 403), while the file and header carry your overall site-wide choice. (Those site-wide signals are published only when you block AI training; an open site publishes none.)

Which AI agents are allowed by default?

Out of the box Agentimus blocks nothing — it's a discovery layer, so every agent is served until you turn on the optional scanner blocking. Even then, an always-allowed list keeps trusted clients flowing: the major search engines (Googlebot, Bingbot, DuckDuckBot, Applebot, Yandex) are recognised automatically and never blocked or flagged, and the AI access tab shows them read-only so you know exactly what's trusted. You can add well-known AI assistants and answer engines (ChatGPT, Claude, Perplexity, …) with one click, or mark any client Allow from the activity review queue. Training crawlers (GPTBot, ClaudeBot, …) are deliberately not on the trust list — those belong to your separate AI-training choice, so trusting them here wouldn't quietly undo an opt-out you may have set.

Can I see if AI is sending me visitors?

Yes — the dashboard's "Traffic from AI" card counts real people who landed on your site from an AI assistant (ChatGPT, Perplexity, Gemini, …), detected from the visit's referrer and the utm_source tag some AI tools add to their links. It's the mirror of the activity log: that shows bots reading your content; this shows AI bringing you readers, with a by-source and top-landing-pages breakdown. Like the rest of the log it's first-party and aggregate-only — no IP, no per-visitor records, nothing sent anywhere. Some AI visits can't be detected (stripped referrers, Google's AI Overviews, cached pages), so read the figure as a floor: at least this many.

Will Agentimus get my site mentioned by ChatGPT or improve my AI rankings?

Honestly: it helps with one half of that, not the other. Agentimus makes your site discoverable and correctly understood — when an AI assistant looks at your site, it can find your content, read a clean version, and describe you accurately. That is what the plugin controls, and it does it well. But whether an AI spontaneously mentions you when someone asks a broad question ("best resources for X") is a matter of authority and reputation — earned over time through genuinely notable content that others reference. No plugin, llms.txt, or schema can manufacture that, and any tool promising "instant AI visibility" is overselling. Agentimus makes sure that when authority does bring an agent to your door, nothing is lost in translation.

Will it slow my site down?

No. The text endpoints are cached and CDN-friendly; there is no front-end JavaScript or CSS for your visitors (the optional, off-by-default WebMCP bridge adds a tiny script only when you enable it, and it stays inert in browsers without the API). The admin app loads only on the plugin's own screen.

Does it expose anything private, or let agents change my site?

No. Agentimus only describes what your site already makes public; it grants no new access. Removing or suppressing an item changes what is advertised, not what is reachable — the underlying endpoints behave exactly as before, behind their own authentication.

Does Agentimus run an MCP server?

Yes — as an opt-in, on WordPress 6.9 or newer. Turn on Settings → Discovery → MCP server and the AI assistants you already use can talk to your site over the Model Context Protocol and run the same permission-checked tools your admin AI gets — twenty-five read-only ones (the ranked findings list, readiness, the pages worth fixing and what each is flagged for, a page’s stored body, the categories and tags your site already uses, the people-vs-machines audience, AI traffic, the request log, edge traffic, search performance and opportunities, Google index answers, bot identification, per-page readability, internal-link suggestions, media search and previews), plus the write tools if you separately allow those (see the next question). It also offers your site's own documents to be read rather than run — llms.txt, llms-full.txt, discovery.json and your agent card are listed as MCP resources, so an assistant attaches them the way it attaches a file, and one you have switched off is never offered. Connecting is usually one approval: give the assistant your server address and it asks you for permission on a page served by your own site, where you choose read-only or read-and-write. Claude, ChatGPT and Cursor work this way today (ChatGPT behind its Developer-mode switch); Codex cannot ask, so it takes a shared token you create on the card and send as a Bearer header. A WordPress application password still works too, for one key per tool tied to a specific user. Nothing becomes public — every request signs in, each tool keeps the same permission checks as the admin screens, and every call is recorded under More → Agent Access. Off by default, and everything needed ships with the plugin.

Can an AI agent write to my site?

Only if you say so, twice. The MCP server starts read-only; a second switch — Let connected agents write — adds the write tools: draft and edit posts and pages complete with categories, tags and a featured image (searched from your media library by title or alt text, or imported from a URL), change one passage of a page rather than replacing it, describe an image for screen readers and image search — the one in your library, or one sitting inside a page, which is a different fix because the page keeps its own copy of the words — set their AI topics and descriptions, and apply Readiness fixes (a fixed list of safe switches that can only turn documented features on, never loosen a protection). Even then, agents can't publish: they leave drafts and pending posts for your review, unless you flip a third switch that allows going live. Writes are held to the site’s own quality bar too: the write tools show agents the same readability rules the in-admin assistant drafts to, and every create or update returns the post’s AI-readability grade. Every write runs as the signed-in user — an agent can never do more than that user could in the editor: filing under existing categories, creating new ones, and uploading images each follow that user's own permissions — and every call is recorded under More → Agent Access, attributed to the key that made it.

Can AI help me write the description, topics and fixes?

Yes, if you're on WordPress 7.0 and have set up an AI provider under Settings → AI. Then Draft with AI appears on the AI description field, Suggest with AI on the Topics field, and Fix with AI on any Readability row that needs work. Agentimus asks your AI through WordPress's shared connectors — it never sees or stores your API key, and nothing is sent anywhere if you haven't set a provider up (the buttons simply don't appear). Every suggestion arrives as ordinary editable text in the field: you read it, change it, and save the post yourself. Nothing is written for you.

Can AI write a whole post for me?

Yes — the writing assistant (the spark button on Agentimus's own screens) turns a described idea into a complete draft. It proposes an outline you can edit first, then writes the title, body, AI description and topics with suggested categories and tags, and shows you everything before a single thing is saved. Create draft opens the post in the editor, where image placeholders arrive with their alt text ready — fill them from your library, or generate them with AI. Revising an existing post happens in the editor instead, where Ask AI works at three scopes: the block you are in, several blocks you select, or the whole post — the last returning a list of proposed edits, each with its reason, that you accept or reject one at a time. It needs the Let connected agents write switch on and an AI provider under Settings → AI, and it never publishes: drafts and pending review only.

Do citation checks use the AI provider I set up in WordPress?

No — it needs its own API keys, and that's on purpose. A visibility check is graded on the sources each engine cited, and WordPress's shared connectors hand back only the answer text; the list of cited sources is dropped before Agentimus could read it. Reading those sources means talking to each engine's own API, so citation checks keep their own keys (Visibility → Citations). They stay on your server and are used for nothing else.

How do I make my plugin appear in the discovery document?

Add a single optional action — no dependency, no library. If Agentimus isn't installed the hook simply never fires: add_action( 'wpdiscovery_register', function ( $registry ) { $registry->register( array( 'id' => 'acme', 'title' => 'Acme', 'type' => 'commerce' ) ); } ); Agentimus also fires the product-aliased agentimus_register; you may hook either. See examples/integrate-your-plugin.php for the full resource schema (capabilities, endpoints, auth, agent cards, MCP tools).

Which hooks can my plugin use?

Registration is a single action, but Agentimus exposes more for deeper integrations, grouped by stability:

  • Stable — frozen at WP_Discovery spec 1.0; build on these: the wpdiscovery_register action with its $registry->register() / add_well_known() API, plus agentimus_entity_types and the agentimus_cache_flushed action.
  • Extension — supported output-shaping filters (signatures may evolve between releases): tune the discovery document, MCP/agent surfaces, llms.txt, schema.org, sitemap, REST discovery and security.txt — e.g. agentimus_envelope, agentimus_documents, agentimus_mcp, agentimus_agent_skills, agentimus_well_known_routed, agentimus_post_types, agentimus_security_txt.
  • Internal — advanced site-owner tuning (Guard, Classifier, Activity, Settings); not a third-party integration surface.
Every hook, with its signature and tier, is catalogued in examples/all-hooks-reference.php.

Is the discovery format an open standard I can read?

Yes. The discovery document implements the WP_Discovery Protocol, an openly-licensed (CC BY 4.0) specification — not a format private to this plugin. Read the spec, the 1.0 JSON Schema and worked examples at https://heera.github.io/wp-discovery-protocol/ (source and conformance tests: https://github.com/heera/wp-discovery-protocol). Agentimus is its reference implementation.

Does Agentimus store IP addresses?

Not by default. The agent-activity log records who fetched your endpoints with no IP addresses at all. One opt-in setting — Store IP addresses for flagged clients — records the IP only of clients flagged as impersonating or spoofed crawlers (never ordinary traffic), so you can block them at your host or CDN. When on, those IPs are personal data: stored on your own site only, kept for a short retention period, cleared when you clear the activity log, and deleted if you turn the setting back off. Nothing is ever sent off your server. If you enable it, disclose it in your privacy policy (Agentimus adds suggested text at Settings → Privacy).

Is the admin interface built from source I can inspect?

Yes — there is no minified-only code. The admin interface is built from Vue 3 source in resources/ with Vite; the source and vite.config.js ship in this package and also live in the public repository at https://github.com/heera/agentimus . Run npm install && npm run build to regenerate assets/admin/ from source.

更新日志:

1.46.0 1.45.0 1.44.0 1.43.0 1.42.0 1.41.0 1.40.1 1.40.0 1.39.0