= 1.1.1 =
* Fix: Backslashes were missing from Markdown when saving to cache – YAML header with quotes or multi-line description was then invalid (“rnrn”, unescaped quotes). Cache is regenerated automatically after update, pages saved in Markdown editor need to be saved again
* Fix: tel:, mailto: and sms: links are no longer rewritten to the website address; anchors (#section) and relative paths refer to the correct page; auxiliary parameter _tmcb is not included in pagination links
* Fix: The converter separates blocks (header, section, FAQ in details/summary) with an empty line, titles and questions are not glued to other text; preserves spaces for line elements; empty icons no longer create lonely “**”; list items preserve links, formatting and nested lists
* Fix: Cloudflare protected emails are decoded even from antispambot(); encoding HTML entities in the title and description in the YAML header are decoded
* Markdown sitemap and llms-full.txt now also contain their own content types (services, portfolio, FAQ…), each in its own section starting from the newest; in the settings, you can limit the number of items for all types and for each separately
* New setting “Max. cache age” (default 30 days) – older pages are regenerated in the background when displayed; the cache is also cleared when changing the template and updating the plugin, locked pages remain
* llms.txt is served dynamically in UTF-8 encoding instead of a physical file in the root of the website (the old file is deleted when saving the editor); without its own content, the default text is displayed instead of a 404 error
* Developer file .claude/settings.local.json removed from the installation package, the update will delete it from the website
* Metabox notifies when the rendered page failed to download; in the settings, a note to the Accept header for websites with full-page cache
* Fix: Backslashes were missing from Markdown when saving to cache – YAML header with quotes or multi-line description was then invalid (“rnrn”, unescaped quotes). Cache is regenerated automatically after update, pages saved in Markdown editor need to be saved again
* Fix: tel:, mailto: and sms: links are no longer rewritten to the website address; anchors (#section) and relative paths refer to the correct page; auxiliary parameter _tmcb is not included in pagination links
* Fix: The converter separates blocks (header, section, FAQ in details/summary) with an empty line, titles and questions are not glued to other text; preserves spaces for line elements; empty icons no longer create lonely “**”; list items preserve links, formatting and nested lists
* Fix: Cloudflare protected emails are decoded even from antispambot(); encoding HTML entities in the title and description in the YAML header are decoded
* Markdown sitemap and llms-full.txt now also contain their own content types (services, portfolio, FAQ…), each in its own section starting from the newest; in the settings, you can limit the number of items for all types and for each separately
* New setting “Max. cache age” (default 30 days) – older pages are regenerated in the background when displayed; the cache is also cleared when changing the template and updating the plugin, locked pages remain
* llms.txt is served dynamically in UTF-8 encoding instead of a physical file in the root of the website (the old file is deleted when saving the editor); without its own content, the default text is displayed instead of a 404 error
* Developer file .claude/settings.local.json removed from the installation package, the update will delete it from the website
* Metabox notifies when the rendered page failed to download; in the settings, a note to the Accept header for websites with full-page cache
= 1.1.0 =
* Fix: Markdown version of the page lost the main content for templates that have a class containing "sidebar" on the page cover (e.g. sidebars1, has-sidebar, no-sidebar) – then only product data and attributes remained in the output
* Fix: Markdown generation will now be protected by the Password Protected plugin (typically on staging websites) – the plugin's own request is verified with a secret token
* The page/product metabox will notify if the rendered page could not be downloaded (login page, HTTP error, empty response) and Markdown contains only basic content
* New files for AI agents: llms-full.txt (extended index with machine-readable endpoints), agents.md (operational manual for AI agents with an editor in the administration) and sitemap sitemap_agentic_discovery.xml included in the WordPress, Yoast and Rank Math sitemap index
* The Markdown version of the page is also available via the .md extension in URL (/stranka.md, /kategorie/index.md)
* X-Markdown-Tokens and Content-Signal headers for Markdown responses (compatible with Cloudflare Markdown for Agents), Content-Signal settings in the administration
* llms.txt according to the v2 specification – rel="describedby" link in the page header and HTTP Link header, links to llms-full.txt and agents.md in robots.txt
* Fix: Markdown version of the page lost the main content for templates that have a class containing "sidebar" on the page cover (e.g. sidebars1, has-sidebar, no-sidebar) – then only product data and attributes remained in the output
* Fix: Markdown generation will now be protected by the Password Protected plugin (typically on staging websites) – the plugin's own request is verified with a secret token
* The page/product metabox will notify if the rendered page could not be downloaded (login page, HTTP error, empty response) and Markdown contains only basic content
* New files for AI agents: llms-full.txt (extended index with machine-readable endpoints), agents.md (operational manual for AI agents with an editor in the administration) and sitemap sitemap_agentic_discovery.xml included in the WordPress, Yoast and Rank Math sitemap index
* The Markdown version of the page is also available via the .md extension in URL (/stranka.md, /kategorie/index.md)
* X-Markdown-Tokens and Content-Signal headers for Markdown responses (compatible with Cloudflare Markdown for Agents), Content-Signal settings in the administration
* llms.txt according to the v2 specification – rel="describedby" link in the page header and HTTP Link header, links to llms-full.txt and agents.md in robots.txt
= 1.0.5 =
* Fix: Sitemap page in administration crashed on PHP memory limit on large websites – content is now generated only once and in batches with memory release
* Expanded AI bot detection from 24 to 56 – added SearchGPT, Microsoft Copilot, Applebot, Meta, CCBot, Mistral, DeepSeek, Perplexity and more
* New categories in access log: Search engines (Yandex, Seznam, DuckDuckGo…) and SEO crawlers (Ahrefs, Semrush…)
* Titles with special characters (colons, quotes, gridlines) no longer break the YAML header of Markdown output
* Minimum required PHP version increased to 8.1
* Fix: Sitemap page in administration crashed on PHP memory limit on large websites – content is now generated only once and in batches with memory release
* Expanded AI bot detection from 24 to 56 – added SearchGPT, Microsoft Copilot, Applebot, Meta, CCBot, Mistral, DeepSeek, Perplexity and more
* New categories in access log: Search engines (Yandex, Seznam, DuckDuckGo…) and SEO crawlers (Ahrefs, Semrush…)
* Titles with special characters (colons, quotes, gridlines) no longer break the YAML header of Markdown output
* Minimum required PHP version increased to 8.1
= 1.0.4 =
* Content negotiation – plugin now returns Markdown when requested with Accept: text/markdown header (common browsers are not affected)
* Link HTTP response headers (RFC 8288) – link to Markdown version and llms.txt in HTTP response header for better AI discovery
* Content Signals in robots.txt – declaration of preferences for content usage by artificial intelligence (ai-train, search, ai-input)
* Protection from cache plugins – Markdown response is no longer cached (WP Rocket, WP Fastest Cache, LiteSpeed Cache, etc.)
* New "Protected words" setting – option to add custom expressions (e.g. mAh, kWh) that should not be split when generating Markdown
* Content negotiation – plugin now returns Markdown when requested with Accept: text/markdown header (common browsers are not affected)
* Link HTTP response headers (RFC 8288) – link to Markdown version and llms.txt in HTTP response header for better AI discovery
* Content Signals in robots.txt – declaration of preferences for content usage by artificial intelligence (ai-train, search, ai-input)
* Protection from cache plugins – Markdown response is no longer cached (WP Rocket, WP Fastest Cache, LiteSpeed Cache, etc.)
* New "Protected words" setting – option to add custom expressions (e.g. mAh, kWh) that should not be split when generating Markdown
= 1.0.3 =
* Fix: Fatal error when saving ACF options page (plugin passed invalid parameter type)
* Updated Toret plugins page in administration
* Fix: Fatal error when saving ACF options page (plugin passed invalid parameter type)
* Updated Toret plugins page in administration
= 1.0.1 =
* New link in admin bar – quick access to Markdown version of current page directly from frontend toolbar
* Dropdown menu with option to open Markdown preview (new window) or go to editor
* New setting "Link in admin bar" – disabled by default, can be enabled in plugin settings
* New button "Regenerate missing cache" – generates Markdown only for pages that do not yet have cache
* Added Czech translations for new strings
* New link in admin bar – quick access to Markdown version of current page directly from frontend toolbar
* Dropdown menu with option to open Markdown preview (new window) or go to editor
* New setting "Link in admin bar" – disabled by default, can be enabled in plugin settings
* New button "Regenerate missing cache" – generates Markdown only for pages that do not yet have cache
* Added Czech translations for new strings