Description
AI agents and AI crawlers are increasingly the ones visiting your site — ChatGPT, Claude, Perplexity, Gemini and their bots (GPTBot, OAI-SearchBot, ClaudeBot, PerplexityBot, anthropic-ai, GoogleBot-Extended) research, recommend, and act on behalf of users. Traditional SEO doesn’t cover this. OARS is AI readiness (sometimes called AI SEO, Generative Engine Optimization / GEO, or Answer Engine Optimization / AEO): it publishes the machine-readable signals AI systems need to find, read, and correctly represent you.
The Open Agent Readiness Standard (OARS) by Knov.ai defines what makes a site machine-operable. This plugin handles the low-hanging fruit automatically and scores the rest.
Free
- oars.json — a machine-readable knowledge declaration served at your domain root, auto-populated from your site title, tagline, categories, and pages.
- llms.txt — a clean Markdown summary for LLMs (guaranteed to meet the 200-word minimum).
- AI bot directives — adds rules for GPTBot, OAI-SearchBot, ClaudeBot, PerplexityBot, anthropic-ai and GoogleBot-Extended to robots.txt.
- Schema.org identity — injects Organization JSON-LD on your homepage.
- Content-Signal headers — declares your AI usage preferences (search / input / training).
- Live readiness score — the scorecard self-fetches your real URLs and headers, so it reflects what agents actually see; every gap is flagged with a fix and a proof link.
- Level 0 baseline — security headers, optional report-only CSP, Link headers, and CORS on your agent files.
- /llms-full.txt, /.well-known/oars.json, sitemap fallback and a guided first-run setup wizard.
- Bring-your-own LLM (optional) — connect OpenAI, Anthropic, or Google to enrich your identity fields and llms.txt narrative.
Pro (separate add-on)
OARS Pro is a separate plugin you install alongside this one (available from tylerewillis.com). It adds:
- Agent API (Level 3) — read-only REST endpoints, OpenAPI 3.1 contract, and an A2A agent card.
- Web Bot Auth signing-key directory and Agent Skills discovery (.well-known).
- Level 4a machine-readable product catalog from WooCommerce.
- Full Level 1–2 scoring with per-check remediation.
- Weekly monitoring with an email alert if your score drops.
- Advanced Schema.org — Service, Product, and FAQPage auto-generated from your pages and WooCommerce catalog.
- Markdown content negotiation (
Accept: text/markdown) for agents. - Scheduled regeneration as your content changes.
- WP Multisite / network management.
- One-click submission to the Knov.ai directory + verification ID.
- OARS compliance badge shortcode.
Pro is licensed per site or per agency. Learn more at tylerewillis.com/tools/oars-for-wordpress.
External services
The OARS Agent Readiness plugin works fully without contacting any third party. Some optional features connect to external services only when you enable them:
- AI enrichment (optional). If you add an API key for OpenAI, Anthropic, or Google, the plugin sends the text content you ask it to process (e.g. your page titles and excerpts) to that provider to generate llms.txt narrative, identity suggestions, or FAQ content. Data is sent only when you click a generate button. Review each provider’s terms and privacy policy: OpenAI (https://openai.com/policies/), Anthropic (https://www.anthropic.com/legal/privacy), Google (https://policies.google.com/privacy).
- Readiness self-check. The dashboard score fetches your own site’s public URLs (e.g. /llms.txt, /oars.json, your homepage) to verify what agents actually receive. These requests go to your own domain, not a third party.
(The optional OARS Pro add-on validates a license key with tylerewillis.com; that disclosure ships with the Pro plugin, not this free plugin.)
The “Scan headers”, “PageSpeed”, and “Validate schema” buttons simply open external tools in a new browser tab; OARS does not send your data to them from the server.
Screenshots
Installation
- Upload the
oarsfolder to/wp-content/plugins/, or install through the Plugins screen. - Activate the plugin.
- Go to Settings OARS to review your readiness score and fine-tune your identity.
FAQ
-
Do I need an API key?
-
No. All Level 1–2 generation is deterministic and works out of the box. An LLM key is optional (for richer narrative text), and a Knov.ai license key unlocks Pro features.
-
Will this overwrite my existing robots.txt or oars.json?
-
No. The files are served virtually by WordPress. If you have a physical robots.txt or oars.json at your web root, your server serves that first and the plugin stays out of the way.
-
Does it work with WooCommerce?
-
Yes — Pro auto-generates Product schema and structured offerings from your WooCommerce catalog.
-
Is this AI SEO / GEO / AEO?
-
Yes — those are different names for the same goal: being discoverable and correctly understood by AI. “AI SEO”, “Generative Engine Optimization (GEO)” and “Answer Engine Optimization (AEO)” all describe optimizing for AI agents and answer engines instead of (or alongside) classic search. OARS implements the concrete, standardized signals behind it — llms.txt, oars.json, AI bot directives, Content-Signal, and Schema.org — and scores your readiness.
-
Which AI systems does it work with?
-
The directives target the major AI crawlers by name — GPTBot and OAI-SearchBot (ChatGPT / OpenAI), ClaudeBot and anthropic-ai (Claude / Anthropic), PerplexityBot (Perplexity), and GoogleBot-Extended (Gemini / Google). The llms.txt, oars.json, and Schema.org output is consumed by any AI agent or LLM, not just these.
-
What data leaves my site?
-
Nothing, unless you opt in. See “External services” below.
Reviews
There are no reviews for this plugin.
Contributors & Developers
“OARS Agent Readiness” is open source software. The following people have contributed to this plugin.
Contributors
Translate “OARS Agent Readiness” into your language.
Interested in development?
Browse the code, check out the SVN repository, or subscribe to the development log by RSS.
Changelog
1.0.0
- First public release. Automates OARS Levels 0–2 — oars.json, llms.txt, AI bot directives, Schema.org identity, Content-Signal + baseline security headers, sitemap, and a live readiness score with per-check fixes and a guided setup wizard. Optional AI enrichment (bring your own key). Higher levels available via the separate OARS Pro add-on.
0.4.0
- Live verification: the scorer now self-fetches your live URLs and headers (robots.txt, llms.txt, oars.json, homepage JSON-LD) so the score reflects what agents actually see — not just which toggles are on.
- oars.json schema-conformance check against the canonical OARS manifest schema.
- Content-Security-Policy support (report-only by default) for Level 0, plus an HSTS check.
- Link headers (RFC 8288), CORS on agent-readable files, /.well-known/oars.json alias, and /llms-full.txt.
- Pro: Web Bot Auth signing-key directory at /.well-known/http-message-signatures-directory (Ed25519 JWKS).
- Pro: Agent Skills discovery (/.well-known/agent-skills/index.json + SKILL.md) and per-offering-page Schema.org.
- New first-run setup wizard, plus “Assess” and “Get verified” links into the Knov.ai directory funnel.
0.3.0
- Pro: Agent API (Level 3) — read-only REST endpoints (/profile, /content, /offerings), an OpenAPI 3.1 contract at /openapi.json, and an A2A agent card at /.well-known/agent-card.json.
- Pro: Level 4a product catalog — a machine-readable WooCommerce feed (pricing, availability, currency) at /oars-products.json.
- Pro: per-level dashboard tabs (Levels 0–4) with a build roadmap + implementor link for Level 5.
- AI Assist: generate/populate identity fields and the llms.txt narrative from your own content (with a connected LLM).
- Compliance level is now grounded against your actual checks (suggests the level you’ve achieved).
- Fixed multi-tab settings save wiping fields on other tabs.
0.2.0
- Added Level 0 (Reachable) checks + baseline security headers.
- Fixed virtual file serving with real rewrite rules so /oars.json and /llms.txt reliably resolve (no more 404s).
- Added XML sitemap detection + fallback generation at /sitemap.xml.
- Scorecard now shows a detail line and a quick link for every check.
- AI enrichment model field now accepts the exact model name (empty = provider default).
0.1.0
- Initial release: Level 1–2 free generation + scoring, optional LLM enrichment, and Pro scaffolding.