RankReady – AI SEO, Schema, llms.txt, AEO and GEO for ChatGPT, Gemini and Perplexity

1.4.1, 2026-10-12

• Fix: API Keys buttons now say how to fix a missing key instead of just reporting it.
• Fix: the browser tab title names the current RankReady tab.
• Fix: saving Brand Identity (including with Enter) now shows a Saved notice.
• Fix: Diagnostics “Copy report” shows a Copied confirmation, and the admin tab bar scrolls on phones.
• Fix: Markdown FAQ answers no longer show & for an ampersand.
• New installs keep same-URL Markdown negotiation off, and Diagnostics warns when a caching layer is detected while it is on. Existing sites are unchanged.
• On-screen WebMCP ability names now match the registered IDs.
• Setup wizard: “Select recommended” wording and WebMCP hint now match the real defaults.
• Pro custom fields: unticking a post type keeps your field choices, enabling one with no field shows a notice, and group sub-field labels are printed.
• Pro headless: non-public post IDs return the same 404 as unknown IDs, and webhook URLs pointing at private or loopback hosts are refused.
• Discovery files send consistent cache headers.
• Bulk Author Changer no longer counts posts the target author already owns and shows validation inline.
• FAQ count clamps to 3-10.
• Robots.txt screen: radios have labels and the state header stays visible.

• OKF index no longer prints an empty dash for entries without a description.

• Fixed: Server-rendered blocks, such as Latest Posts, reusable blocks and ACF blocks, now appear in your Markdown output and llms-full.txt.

• Fixed: Cached Markdown now refreshes when a summary, FAQ, term or registered custom field changes, including on sites with Redis or Memcached.
• Improved: Markdown cache hits no longer re-render builder pages.
• New: The rankready_md_watched_meta_keys filter lets add-ons register custom fields that should refresh the Markdown cache.
• New: The rankready_md_cache_salt filter lets add-ons refresh all cached Markdown when one of their settings changes.
• Security: Dismissing the nginx notice now checks the user’s capability.
• New: The dashboard now links to the HostMyBlog Community and to email support at rankready@hostmy.blog.
• Improved: The Privacy Policy guide in WordPress now has suggested text for the AI crawler log and referral counts. Neither stores IP addresses, user agents or cookies.
• Improved: The plugin description is shorter, so WordPress.org no longer cuts it off.
• Fixed: Markdown pages now send the same Content-Signal header as your robots.txt.
• Fixed: Pages that contain only RankReady blocks now return their Markdown instead of an error.
• Fixed: Markdown author details join topics with commas and print the bio safely.
• Fixed: The FAQ and Summary buttons in the Elementor editor load again and show a clear message when an AI key or post type is missing.
• Improved: The FAQ, Summary and Author Box blocks explain in the editor why they are empty and when they are switched off.
• Improved: Settings screens confirm “Settings saved.” next to the Save button. Updated admin scripts no longer get stuck on a cached copy.
• Improved: Diagnostics skip password-protected posts, and the Plugins screen shows a Manage License link when Pro is licensed.

1.4.0, 2026-09-30

• New: Every setting links to the help article that explains it.
• New: You can see what each Pro setting does before you buy. Writing summaries and FAQs one post at a time stays free, with no limit.
• New: A short setup message now tells you exactly what is missing, such as a brand name or an AI key.
• Improved: Admin notices are shorter, close with one click, and stay off the Upgrade screen.
• Improved: You can leave the setup wizard from any step, and the last step points to what still needs doing.
• Improved: Screens are easier to read on a phone, with clearer links, checkboxes and keyboard focus.
• Improved: AI model lists are checked against each provider, and old model names are moved to the current one for you.
• Improved: “Enable AI Summary” and “Enable AI FAQ” are now called “Show on frontend”, which is what they always did. Your saved settings are untouched.
• Fixed: Smoother updates from older versions.
• Fixed: Turning off llms.txt, Markdown or reviewer details now switches them off completely.
• Fixed: Making a summary or FAQ from several places at once no longer creates duplicates.
• Fixed: Deleting the plugin on a multisite network now cleans every site that opted in.
• Fixed: Wording across the plugin now says only what RankReady actually does.
• Fixed: Many small display fixes in the dashboard, the post editor and the setup wizard.

1.3.1, 2026-09-01

• New: Cloudflare has its own Settings subtab (API Keys → Cloudflare → Advanced). The connect form is always available, with a warning when Cloudflare is not detected, so staging and DNS-only sites can still connect.
• Fixed: OpenAI “Verify Key” no longer fails on GPT-5.x with “max_tokens or model output limit was reached” — the probe now allows enough completion tokens for reasoning models.
• Improved: LLM model dropdowns load live model lists from each provider when an API key is saved; a small offline fallback is used only when no key is set or the fetch fails.
• Improved: Model dropdown labels use exact provider model IDs (e.g. claude-sonnet-5) so each option is unambiguous; tier guidance remains in the field description below.
• New: “Refresh list” button next to each model dropdown fetches the latest models on demand without waiting for the cache to expire. A successful “Verify Key” also refreshes that provider’s list.
• Improved: “Verify Key” sits inline beside each API key field (all four LLM providers and DataForSEO).
• Fixed: OpenAI model selection no longer reverts after save — removed the hardcoded GPT-4o allowlist that rejected newer model IDs.
• Improved: OpenAI model list hides chat-only variants (IDs containing -chat). Gemini hides non-text models (robotics, TTS, image, transcribe, computer-use, and similar).
• Improved: If your saved model is retired and missing from the live list, it stays visible with a deprecated notice so you can pick a replacement.
• New: Delete generated Summary or FAQ from the post editor metabox — an inline danger link in the status line (e.g. “Summary generated 8 minutes ago. Delete summary”).
• Improved: Regenerate cooldown countdown shows in the status line (“You can regenerate again in 59s.”) instead of changing the button label to “Wait 59s”.
• New: llms.txt setting to use .md URLs in post links (on by default when Markdown endpoints are enabled). Hides the redundant “Append .md to any page URL” hint when active.
• Improved: llms-full.txt Source: lines use the same .md URLs when that setting is on.
• New: “Clear cache” button on the llms.txt settings screen rebuilds /llms.txt and /llms-full.txt on demand and purges CDN/page-cache layers for those endpoints.
• Removed: Dashboard YouTube walkthrough embed (no third-party video in wp-admin).

1.3.0, 2026-08-24

• Improved: Settings are reorganized into focused tabs, Dashboard, AI Visibility, AI Content, Insights, and Settings, with subtabs so Brand Identity, robots.txt, llms.txt, Markdown, WebMCP, OKF, Summary, FAQ, Author Box, and Schema are easier to find.
• Improved: The post-edit RankReady UI is three metaboxes (AI Summary, AI FAQ, AI Visibility) instead of one combined box. Each box only appears on post types that use that feature, and saves are isolated so hiding a box cannot clear another feature’s settings.
• New: Generate Summary and Generate FAQ right in the post editor, no block or widget required. Same REST endpoint and 60-second cooldown as the block.
• New: Classic Editor shortcodes [rankready_summary], [rankready_faq], and [rankready_author]. If a matching block or shortcode is already in the post, auto-display stays out of the way.
• New: Enable toggles for HTML summaries and FAQs, plus a single auto-display placement control (before / after / both / off). Turn the on-page box off and the generated text still feeds Markdown, OKF, and WebMCP.
• New: Homepage and blog-index Markdown as their own surfaces (/index.md and your Posts page .md), on by default and independent of the Pages post type.
• New: Site-wide AI Snippet default (max-snippet:-1) on AI Visibility → robots.txt, with the same control on the dashboard. Each post can still override it.
• Improved: Summaries, FAQs, and the Author Box only output on the post types you selected, including schema, Markdown, and WebMCP.
• Improved: Dashboard AI Content tiles show Disabled/Off when no post types are set, and counts only published posts of those types. The posts-list RankReady column spells out Summary/FAQ and flags posts excluded from AI.
• Improved: In the block editor, RankReady metaboxes start collapsed and follow the document sidebar. Classic Editor is unchanged.
• Improved: Physical robots.txt stays in sync on the first save of crawler, llms/Markdown, and Content Signals settings, not only later updates. RankReady no longer writes an empty managed block when crawler rules and Content Signals are both off, and deactivation strips the same block formats as a re-sync.
• Fixed: Turning off llms.txt, Markdown, or OKF no longer leaves a raw 404 Not Found on those URLs. Stale rewrite rules are cleared so WordPress handles the request again (llms-full.txt follows the master llms.txt toggle).
• Fixed: Exclude from AI applies to Markdown and WebMCP as well as llms.txt and OKF.
• Fixed: WebMCP is off until you turn it on (including skip-onboarding). Sites that already saved it on are unchanged.
• Fixed: The setup wizard will not show Congratulations until setup is actually finished.
• Several bug fixes and stability improvements.

1.2.1, 2026-07-20

• New: Block AI crawlers, a “Block Crawlers” list on the AI Crawlers tab adds Disallow rules to robots.txt for the bots you choose. Off by default, so your robots.txt is unchanged until you use it.
• New: Connect Cloudflare with a scoped API token and RankReady creates the Markdown cache-bypass rule for you automatically.
• New: Googlebot and Facebook’s link crawler now get their own robots.txt entry that repeats the rules your site already applies. They are listed for tools that check crawler access, and what they are allowed to reach does not change.
• Improved: AI Summaries and FAQs now generate in your site’s language instead of English (a German site gets German content), section headings included; question research uses your language and country too.
• Improved: llms.txt, Markdown, WebMCP and OKF endpoints resolve more reliably across server and SEO-plugin setups, with clearer nginx guidance when /.well-known/ is blocked.
• Fixed: robots.txt no longer collects a duplicate Content-Signal line every time you save settings. Any duplicates already in the file are removed on the next save.
• Several bug fixes and stability improvements.

1.2.0, 2026-07-09

• Added: AI SEO tips by email, opt in from the setup wizard or the dashboard for practical AI SEO tips and product updates. Optional and one-time; nothing is sent unless you tick the box, and it never appears again once you subscribe. See “External services”.
• Fixed: AI provider keys (OpenAI, Claude, Gemini, DeepSeek) now save exactly what you enter, every time.
• Fixed: the “Saved” confirmation now reflects the real result, if a save cannot complete, you are prompted to reload instead of seeing a false success.
• Fixed: Markdown (.md) versions of your pages stay out of Google’s index even when served from a page cache (WP Rocket, LiteSpeed, W3 Total Cache, WP Super Cache). CDN and browser caching are unaffected.
• Fixed: no more duplicate robots meta tag on sites without a third-party SEO plugin, and the snippet-preview control now applies correctly alongside All in One SEO.
• Fixed: WebMCP resource toggles now match what is actually served, public content is available to AI agents out of the box and stays enabled after saving, while private resources stay off until you enable them.
• Improved: llms.txt and llms-full.txt stay out of search results while remaining fully readable by AI agents.
• Improved: Nginx sites that block /.well-known/mcp.json now get the exact one-line fix inside Diagnostics.
• Improved: serving Markdown to AI agents that request it now works out of the box, including on Cloudflare, with no effect on site speed.
• Improved: Open Knowledge Format settings and bundle download are now in one card, and manual cache-flushing is gone, the cache refreshes automatically when you publish or edit.
• Fixed (security): password-protected posts are never included in the OKF bundle (/okf/), they were previously readable there without the password.
• Fixed (security): the WebMCP author lookup now returns only authors with published content, preventing anonymous user enumeration.
• Fixed: AI Summary and FAQ generation on OpenAI (the request now uses the parameter current GPT models require, so generation no longer fails).
• Fixed: “Verify Key” for OpenAI now runs a real generation check, so it can no longer report success while generation would fail.
• Fixed: the setup-wizard email opt-in is unchecked by default (explicit opt-in).
• Improved (hardening): AI-crawler log flood throttle collapses by path so query-string noise can’t bypass it; “Delete all data” now removes every RankReady option; the WebMCP manifest advertises only resources with a real ability; internal HTTPS probes verify TLS by default.

1.1.2, 2026-06-17

• Google Open Knowledge Format, serve an AI-readable OKF bundle of your content at /okf/. One click, auto-synced on publish.
• Several bug fixes and improvements for a better experience.

1.1.1, 2026-06-02

• Fixed: API keys (OpenAI, Claude, Gemini, DeepSeek, DataForSEO) would not save on first entry, the key verified but saved blank. The first save now stores it correctly.
• Fixed: pages could show raw Markdown to visitors behind Cloudflare APO or other caches that ignore Vary: Accept. Markdown is now served only at the distinct .md URLs (e.g. /post-slug.md, /index.md), which are cache-safe; same-URL negotiation is an opt-in toggle.
• Fixed: updating the plugin no longer resets your AI Summary post-type / CPT selection back to the default, the onboarding step only seeds defaults for settings that were never saved.
• Added: full multilingual support, Turkish, CJK, Arabic, Hindi, Cyrillic and other non-Latin scripts render correctly in Summaries, FAQs, and Author profiles (stored as real UTF-8). One-time silent migration of existing content.
• Added: Squirrly SEO compatibility, AI schema merges into Squirrly’s JSON-LD graph instead of emitting a duplicate block.
• Added: SWIS Performance compatibility, shows the exact wp-config exclusion snippet so AI endpoints (llms.txt, .md, mcp.json) stay fresh.
• Added: EWWW Image Optimizer detected (images only, no conflict with RankReady’s text endpoints).

1.1.0, 2026-06-01

• Consistent block/widget names and a dedicated “RankReady” group in Gutenberg and Elementor; smart Generate/Regenerate button on any post type.
• All four AI providers (OpenAI, Claude, Gemini, DeepSeek) detected everywhere, with automatic migration of retired model IDs.
• Lighter front end, assets load only on pages using a RankReady block or widget; no front-end JavaScript. No data loss on update.

1.0.1, 2026-05-27

• Fixed homepage Markdown URL on static-front-page sites (was emitting example.com.md; now /index.md).
• Fixed AI Summary settings not saving (settings-group mismatch).
• WP-Cron diagnostic now accepts external system cron (no false warnings on managed hosts).
• Cache headers audited to RFC 9110/9111 with CDN content-negotiation fixes and a Cloudflare APO auto-detect notice.
• Removed the extra “Enable” step on togglable cards, tick the toggle and Save.

1.0.0, 2026-05-26

First public release. The AI-search layer for WordPress: unlimited manual AI Summaries and FAQ schema, llms.txt + llms-full.txt, Markdown endpoints, 29 AI-crawler controls with robots.txt sync, E-E-A-T + Article/Speakable schema (coexists with Rank Math / Yoast / AIOSEO without duplicate output), content freshness, Insights, broad cache-plugin compatibility, multilingual llms.txt, and a Diagnostics suite.