| 开发者 | caidance |
|---|---|
| 更新时间 | 2026年7月20日 10:47 |
| PHP版本: | 8.1 及以上 |
| WordPress版本: | 7.0 |
| 版权: | GPLv2 or later |
| 版权网址: | 版权信息 |
Sitemap: line to robots.txt. The sitemap URL is verified live first (your SEO plugin's index, else the WordPress core sitemap — it must load and be real sitemap XML before the fix is even offered). Works in both robots.txt states: a physical file gets the append with the complete original stored for byte-for-byte restore; a WordPress-generated robots.txt gets a standing output line with no file written and instant revert. Never adds a second line next to an existing declaration./wp-content/plugins/caidance-ai-readiness, or install via the WordPress Plugins screen.No. The scanner runs entirely in PHP against your own site. Results are stored in your WordPress database. Nothing is sent to Caidance or any third party.
Only the four fixes above, and only with your approval: it can create an llms.txt file, declare your verified sitemap in robots.txt, remove AI-crawler-only blocks from robots.txt (storing the complete original for one-click restore), and switch on homepage Organization/WebSite JSON-LD output. You preview the exact change first, nothing happens until you click approve, the result is re-checked and recorded in a local evidence log, and one click reverses each. Everything else the plugin does is read-only.
No. Every fix is conflict-aware. llms.txt is create-only — an existing file or a plugin serving one is left alone. robots.txt edits never touch groups that cover non-AI crawlers, and the original file is stored for byte-for-byte restore. The Sitemap fix declares only a sitemap that actually responded when checked, prefers your SEO plugin's own sitemap, and steps back automatically if anything else already declares one. Schema output defers to Yoast SEO, Rank Math, All in One SEO, The SEO Framework, SEOPress, and Slim SEO — and goes silent automatically if you install one later.
No. Stack detection runs locally against your own active-plugins list and a curated table shipped inside the plugin — no remote calls, nothing stored outside your database. The only time anything travels is when YOU click the Alignment Review link: the detected platform names ride along as URL parameters to pre-fill the first questions of the assessment.
The file stays — you approved its creation and it is your site content. If you want it removed, use Revert this fix on the Tools page before uninstalling (revert deletes the file only if it still exactly matches what Caidance wrote).
No. The plugin is fully usable on its own. The links to caidance.ai are optional — they point to the broader off-site analysis the plugin cannot do from inside WordPress.
No. The scan runs only when you click Run scan now or on the weekly schedule. It runs in the admin context, never on front-end pages.
Open Tools → Caidance Scan. Each failed check has a plain-English explanation of why it matters and the first step to fix it. The fixes are ordered by impact, so start at the top.
Traditional SEO targets Google ten-blue-link rankings. AI search is different. AI agents read schema, llms.txt, and structured signals that do not always overlap with classic SEO. A site that ranks well on Google can still be invisible to ChatGPT. This plugin checks the AI-specific signals.
It looks at your robots.txt and checks whether five major AI crawlers are allowed: GPTBot (OpenAI), ClaudeBot (Anthropic), PerplexityBot (Perplexity), OAI-SearchBot (OpenAI search), and Google-Extended (Google AI). Some sites intentionally block these. If yours does, toggle the check off in Settings.
The 10 universal checks all work on a WooCommerce site, and the llms.txt fix is store-aware: it reads your assigned WooCommerce pages so cart, checkout, and account never appear in the file, and it features your shop page plus your top product category and brand pages instead — the pages an AI shopping agent actually needs. A future version will add WooCommerce-specific checks (product schema, breadcrumbs, review aggregates).
Single-site only for v1. Multisite Network support is on the roadmap.
Yes. The scan reads your site the same way an AI crawler would — including any HTML your cache layer serves.
Firewall and CDN bot protection (Cloudflare Bot Fight Mode, for example) sometimes challenges the plugin's own scan requests — the same way it challenges other automated clients. Since 1.4.2 the scanner detects this instead of failing the checks: affected checks are marked Unverified and excluded from your score, a banner explains what was detected, and the weekly drift watch stays quiet until the site is readable again. Check your firewall or CDN bot settings, or allowlist requests from your own server, then re-run the scan. Worth knowing: if the scanner is being challenged, real AI crawlers may be too.