內容簡介
Agentimus 是一款全方位的 AI SEO 外掛，專為 AI 代理時代設計，結合了 AEO（回答引擎優化）和 GEO（生成引擎優化）。它幫助網站提高可讀性和可引用性，並讓 AI 助手如 ChatGPT 更容易找到並引用您的內容。
【主要功能】
• 提高網站可讀性與可引用性
• 內建模型上下文協議（MCP）伺服器
• 一鍵啟用 AI 工具撰寫與編輯文章
• 提供網站分析與優化建議
• 兼容現有的 SEO 外掛功能
外掛標籤
開發者團隊
② 後台搜尋「Agentimus – AI SEO, llms.txt & MCP for AI Agents」→ 直接安裝（推薦）
📦 歷史版本下載
原文外掛簡介
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
Findings — everything open across your site in one ranked list: pages losing a click they already earn, and anything the setup checks caught — ordered by what each costs you, every row carrying the button that lands on the fix. The nav count only measures work you can act on.
Your Content — one row per post, page or product: what it is found for, whether it answers that, and anything else it needs. Set aside anything never meant to be quoted and it stays listed, not silently dropped.
With an SEO plugin — or instead of one
No SEO plugin? You don’t need one. Agentimus covers the search basics: per-page SEO titles, social share cards (Open Graph), canonical links, meta descriptions, and an XML sitemap carrying the last-changed dates core’s own leaves out.
Already running Yoast, Rank Math, SEOPress, AIOSEO or The SEO Framework? It detects them and steps aside on every overlapping surface — titles, cards, schema, sitemap. No duplicate tags: it adds only the AI layer your SEO plugin doesn’t cover.
Operate your site from your AI agent (MCP) — opt-in
A Model Context Protocol server on your own site — one switch runs an MCP server at /wp-json/agentimus/v1/mcp; the library ships with the plugin, nothing extra to install.
Connect by approving, not by pasting keys — an assistant asks you for permission on a consent page on your own site; you choose Read only or Read and write, and each gets its own key and Disconnect. OAuth 2.1 with PKCE, nothing brokered by a third party; a revocable shared token covers clients that can’t ask.
Read your site’s data — connected agents run the read-only tools — score, AI traffic, request log, bot identification, page previews — and your llms.txt, discovery and agent-card documents are offered as readable resources.
Draft, edit and publish posts — behind two more switches — turn on Let connected agents write and an agent can create and edit posts and pages fully dressed (categories, tags, featured image, AI topics and descriptions), apply Readiness fixes, and reuse pictures from your media library. A third switch lets it publish; otherwise it leaves drafts for review.
Safe by construction — every write runs as the signed-in WordPress user, never exceeding their permissions, and is recorded under More → Agent Access. Nothing is public, and with the switch off the write tools do not exist.
Write with AI in wp-admin — the built-in assistant (opt-in)
Idea → draft without leaving wp-admin — a spark button opens the writing assistant: describe what you want, shape the outline it proposes, then preview the complete draft — real blocks, AI description, topics, categories, tags. A page is written as a page: no invented sections. Nothing is saved until you click Create draft, and it never publishes.
Ask AI in the editor — one block, a selection, or the whole post — rewrite or extend the block you’re in, change several with one instruction, or ask about the whole post and get a list of proposed edits, each with its reason, to accept or reject one at a time. Blocks the plan doesn’t name are never touched.
Images where you write — alt-filled placeholders in drafts, Generate image from the alt text on every image block, a Featured image (AI) panel — or pick from your library. Runs on WordPress’s built-in AI Client (7.0+); Agentimus never sees your key, and every AI button hides until a provider is set up.
Tell your team — and announce what you publish
Reports where you already look (opt-in) — send Agentimus’s reports to Telegram, Slack, Discord, a Google Sheet or any address of your own. Choose the moments that matter: a new finding, an unfamiliar bot, the weekly digest. Every service stays yours, and nothing is sent until you switch it on.
Announce a new post (opt-in) — Agentimus offers the words for X and LinkedIn; you approve them or edit them first. Each connection is an app you own, so the post comes from you, not from us. A ledger shows what went out, what is queued and what failed, with the reason.
Control — who may use your content
robots.txt content-signals + AI-training blocklist — declare your content-usage policy and block model-training crawlers (GPTBot, CCBot, ClaudeBot, Google-Extended, Bytespider, …) by name, while leaving read/cite bots free.
Block scanners & scrapers (opt-in hard block) — robots rules are a polite request; this enforces them, returning 403 to the user-agents on your denylist. Your always-allowed list is never blocked: pre-trust well-known assistants with one click, search engines are recognised automatically, and SSL renewals stay reachable.
Exposure controls (opt-in, all OFF by default) — switches that close what stock WordPress reveals to anonymous crawlers: username enumeration, author archives, the WordPress version, the auto-generated
Visibility — who is reading you
Report — what AI did on your site between any two dates: reads, visits from AI answers, and where you stand. Today, a week, or dates you pick; every block says how fresh its numbers are.
Agent activity log — which AI crawlers and agents actually fetch your content and endpoints (GPTBot, Claude, Perplexity, Googlebot, …), recorded first-party, with no IP logging by default.
Activity to review — a nav-bar queue surfaces clients worth a second look, with one-click Block or Allow. Nothing is blocked unless you say so.
Request Log — every recorded request, one row each. Filter by client, endpoint, network, user-agent and date to see what a single bot fetched.
Agent Access — who authenticates and acts: assistants approved, keys created or revoked, abilities run, requests refused. A record, not a guard — it names the key, never the person.
Traffic from AI — the real visitors an AI assistant sent you, day by day, by assistant and landing page — aggregate counts only, never a row for one person, no IP. An opt-in CDN mode keeps counts accurate behind a cache.
Edge traffic (Cloudflare, opt-in) — what Cloudflare answered or blocked before your server saw the request: cache hits and edge blocks no server-side log can see. It warns when the edge disagrees with your policy.
You decide how long it’s kept — retention, nightly auto-delete and a hard size cap, so the log never outgrows your host.
Citation checks (opt-in) — track each brand, product or person you choose across ChatGPT, Perplexity, Gemini and Claude: whether it is mentioned, linked, and how it ranks against rivals. Off by default; bring your own API key (the one feature that makes an outbound request — see External services).
Classic search, measured — Bing & Google (opt-in)
Search Performance & Opportunities — connect Bing Webmaster Tools and/or Google Search Console (a key held on your own server, no third-party proxy) and see what people searched, how often you appeared, and which pages sit one improvement from page one. Every number is the engine’s own, never estimated; probe traffic is named, not blended in.
In the index — Google’s and Bing’s — whether the indexes behind AI Overviews, Gemini, ChatGPT search and Copilot hold your pages: the whole site in rotation, every verdict in the engine’s own words, problems grouped with deep links to the fix, any page re-checked live — plus week-on-week trend, Google Discover, and your registered sitemap’s health.
Content — clean, machine-readable output
Markdown delivery — request any page as clean markdown by appending .md to its URL. An Accept: text/markdown mode also exists.
/llms.txt & /llms-full.txt — an llmstxt.org index of your pages, topics and recent posts, plus a full-text edition an agent ingests in one request.
JSON-LD — WebSite + Person/Organization, plus BlogPosting and BreadcrumbList on posts. Defers to Yoast, Rank Math, SEOPress, AIOSEO and The SEO Framework, so never duplicate schema.
Topics for AI — say what each post is about in plain words; they become the JSON-LD keywords and a line in the page’s .md. Type your own or fill them from tags and categories. Nothing shows on the page.
AI description — a one-line summary per post; it becomes the JSON-LD description, the lead of its .md, and its unless an SEO plugin owns it. Blank uses the excerpt; a sub-switch keeps it out of your
XML sitemap — with no SEO plugin, Agentimus serves your sitemap at /wp-sitemap.xml and advertises it in robots.txt and llms.txt; with one, it links theirs.
Change feed — /agentimus-changes.json lists added, updated and removed pages (with ?since=), so an assistant re-checks only what changed.
Identity & contact
Author / site identity — a profile sentence, expertise topics and linked profiles (sameAs) feed llms.txt and JSON-LD — the highest-signal lines for agent retrieval.
security.txt — optionally publish an RFC 9116 disclosure contact at /.well-known/security.txt.
Readiness report
A one-screen score of how machine-readable your site is, with a plain-English checklist of what’s enabled and what’s still missing.
Agent preview — see the exact JSON-LD and Markdown an AI agent receives for the whole site or any page, then copy it. It shows what would ship even when the feature is off, and a matching preview sits in the post editor.
Readability tips — as you write, a panel flags what makes a page hard to read, section and quote: thin content, missing headings, no opening summary, a nav-heavy page, images without alt text. Most of those serve search engines and screen readers just as much, and the panel says so. Editor-only — nothing shows to visitors.
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:
/.well-known/discovery.json — an owner-curated document describing the site’s identity, capabilities, APIs and agent cards. Other plugins declare themselves through one hook.
/.well-known/agent-card.json and /.well-known/mcp.json — an A2A agent card and an MCP manifest, generated for you.
Standards-aligned .well-known endpoints — an RFC 9727 api-catalog, plus — only when the capability exists — an MCP server card and an Agent Skills index. Response signing (RFC 9421 / Web Bot Auth) uses an Ed25519 key that never leaves your server.
WordPress Abilities API — the same read-only tools are registered as abilities, each gated by the capability of its screen, so WordPress’s built-in AI can read them. An off-by-default switch adds the write abilities.
The plugins you run, described — a WooCommerce store says it sells products; FluentCart, FluentCommunity and Fluent Support say what they hold. Plugins that keep everything behind a login are named as such. Switch off anything you would rather not announce.
Zero-config auto-discovery — reads your REST namespaces, public post types and the Abilities API, so a site is described even when no plugin declares itself. The Discovery Hub shows what an agent sees.
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.
External services
Agentimus makes no outbound requests by default: no remote scripts, fonts or analytics, and the agent-activity log stays in your own database with no IP addresses. (IP storage is optional, off by default — see the FAQ.)
Every outbound feature is opt-in and off by default.
Data sources & IndexNow: connecting Cloudflare, Google Search Console or Bing Webmaster Tools polls that service with your own key. The optional IndexNow switch announces published and removed URLs to search engines via api.indexnow.org (https://www.indexnow.org/) — only the changed URL list is sent.
Verify bot identities makes DNS lookups and, once a day, downloads the IP-range files bot operators publish to verify their crawlers (Google, Microsoft, DuckDuckGo, Apple, OpenAI, Perplexity — or a URL you add yourself). Only those files are fetched; nothing about your site is sent.
The same setting verifies Web Bot Auth signatures — the standard where an AI agent signs its request cryptographically. To check one, Agentimus fetches (and caches) the signing operator’s public key directory at /.well-known/http-message-signatures-directory, only when a signed request arrives, and sends nothing about your site with it. Operators publishing one today:
OpenAI — https://chatgpt.com/.well-known/http-message-signatures-directory · https://openai.com/policies/terms-of-use · https://openai.com/policies/privacy-policy
Google — https://agent.bot.goog/.well-known/http-message-signatures-directory · https://policies.google.com/terms · https://policies.google.com/privacy
Citation checks: when you enable them and add your own API key for a provider, Agentimus sends the prompts you configured to that provider to check whether it mentions and cites your site — only for the engines you turn on, and only when a check runs. Your keys are stored on your own site and used solely for these calls. The providers, with their terms and privacy policies:
OpenAI (ChatGPT) — https://openai.com/policies/terms-of-use · https://openai.com/policies/privacy-policy
Perplexity — https://www.perplexity.ai/hub/legal/terms-of-service · https://www.perplexity.ai/hub/legal/privacy-policy
Google (Gemini) — https://ai.google.dev/gemini-api/terms · https://policies.google.com/privacy
Anthropic (Claude) — https://www.anthropic.com/legal/consumer-terms · https://www.anthropic.com/legal/privacy
URL-like strings in the plugin’s output are labels, not requests — the discovery documents’ $schema value names the format (never fetched), and the example.com URLs in examples/ are documentation placeholders.
