CyberSEO Pro version changelog

Support for Anthropic Claude 3.5 Haiku.

The XML tag attribute values can now be assigned to WordPress custom fields, e.g.: tag:attr->my_custom_field.

The %post_tags% and tegories% placeholders are now available in "Custom fields".

Support for XML tag attributes.

%xml_tags_attr[name][attr]% placeholder.

Improved: Saved image deletion for links starting with "/".

Support for FLUX.1, a cutting-edge text-to-image AI model available through GoAPI that features low moderation and the ability to generate photorealistic images of real people, including celebrities, making it ideal for news illustration.

Support for Stable Diffusion 3.5.

Parser for Amazon product search pages.

The full text extraction process will now use the title of the post from the feed instead of the title from the web page.

New %gpt_prompt[]% placeholder that allows AI-generated content to be used within shortcodes.

Support for Google Gemini 1.5 Flash, Perplexity Llama 3.1 Sonar 70B Online and Perplexity Llama 3.1 Sonar 405B Online models on the "Article generation" tab.

Support for Google Gemini 1.5 Flash, Perplexity Llama 3.1 Sonar 8B Online, Perplexity Llama 3.1 Sonar 70B Online models as the default AI model in General Settings.

The Markdown to HTML algorithm.

TBS support.

Maintenance updates and usability improvements.

CyberSEO Pro Syndicator, Amazon product import, Markdown to HTML converter.

The video search accuracy for the %youtube_video[%post_title%]% placeholder.

Support for Gemini Flash 1.5 in AI Spinner and AI Translator.

Swedish and Slovak language support in AI Translator.

WPML and Polylang support.

The AliExpress CSV (AI) preset now uses the GPT-4o mini model by default and generates rich-formatted HTML ads with much better quality.

An error in the log when importing images from Pixabay.

Support for the new OpenAI o1 mini (both official and OpenRouter API) with a maximum output window of 65,536 tokens.

Now the %gpt_image_prompt% placeholder in the “batch insert” mode of the “Media enrichment” menu generates an individual image prompt for each text block of the article, before which the generated image is inserted.

YouTube channel links are now automatically converted to their corresponding RSS feeds.

Markdown to HTML conversion.

The %youtube_transcript[]% shortcode, which imports the transcript for a given link to a YouTube video.

Support for GPT-4o in spinner and translator.

Syntax highlighting issues in control panel fields using CodeMirror for PHP and HTML code. The problem causing blocked resource loading has been resolved by switching to a more reliable CDN for loading CodeMirror libraries.

Support for the latest image generation models Stable Diffusion 3, Stable Image Core and Stable Image Ultra.

Support for Perplexity Llama 3.1 Sonar 405B Online with internet access in Auto-comments and Article generation.

Support for the new OpenAI's chatgpt-4o-latest model. It is the most powerful and constantly updated version of GPT-4o with a maximum output window of 16,384 tokens.

Support for Meta Llama 3.1 and Mistral Large 2 in Auto-comments, Article generation, Translate and AI Spinner.

Support for importing media attachments from RSS feeds, ensuring better handling of embedded media files.

Integrated AI Translator using the efficient and cost-effective OpenAI GPT-4o mini language model.

Support for the new Meta Llama 3.1 405B model.

Placeholder %gpt_image_prompt%, which analyzes the text of the current article and generates a prompt to create a relevant image using supported graphic models like Midjourney, DALL-E 3, and Stable Diffusion.

Google News support.

The integrated AI Spinner now operates significantly faster and with higher quality. Additionally, it now supports multiple languages, including German, Spanish, Chinese, French, and many more.

The "Auto-comments" generation panel now offers a wide selection of different AI models.

The "AI article generation" panel now offers a wide selection of different AI models.

Support for OpenAI GPT-4o mini.

The "Default AI model" selector in the "General Settings" menu. This model is used by the %gpt_keywords% and %gpt_keyphrase% placeholders. Since these placeholders analyze large amounts of text, the list of supported AI models only includes versions with a cheap API and a large context window, such as OpenAI GPT-4o mini, Google Gemini Flash 1.5, Anthropic Claude 3 Haiku, etc. Thus, the cost of using the language model APIs will be negligible.

Support for the new Google News encrypted link format. Both Web and RSS Google News link formats are supported.

WordPress 6.6.

Support for Claude 3.5 Sonnet has been added. It's now the default model for the [claude] shortcode.

OpenAI GPT-4o is now the default model for the [gpt_article] shortcode.

Section description handling the [gpt_article] shortcode.

PHP 8.2 and higher.

Post meta title generation for the supported SEO plugins: Yoast SEO, Rank Math SEO, The SEO Framework, SEOPress.

"Convert PNG images to" preference to automatically convert locally saved PNG images to JPEG or WebP format.

Add images to Media Library.

Generation of article headings for [gpt_article].

Support for OpenAI GPT-4o.

OpenAI GPT-4o is now the default model for the [gpt_article] shortcode.

The "sectionN" parameter is now supported in the [gpt_article] shortcode.

The use of the "structure" parameter in the [gpt_article] shortcode. In previous versions, it was erroneously added to the prompt that generated the text sections of the article. Now it is only used when generating headings, as described in the documentation.

Scrolling of the browser window when clicking on an Amazon image thumbnail.

Atom feeds handling.

Google Gemini and Antrhpic Claude text generation routines.

The [gpt_article] shortcode now uses the latest Geming Pro 1.5 model when using "google-gemini-pro" as the engine parameter.

The behavior of the [gpt_article] shortcode when used to generate a 1-section article.

"Post slug" and "Remove emoticons from post slugs" to the "Post template" tab.

Categories and tags are now assigned after post-translation.

Problem with filtering words containing the "&" symbol.

An issue with exporting presets.

PHP 8.2.

"Batch insert" and "Replace all images" options in the "Media enrichment" menu.

The HTML content import routine.

The post insertion routine.

The "Protect post titles" setting. Should be used to prevent titles from being synonymized when importing from CSV.

%gpt_keywords%, %gpt_keywords_en%, %gpt_keyphrase% and %gpt_keyphrase_en% processing.

Anthropic Claude 3 models "claude-3-opus-20240229" and "claude-3-sonnet-20240229" now available for [claude] shortcode.

Midjourney support via goapi.ai.

[midjourney] shortcode.

"Replace fisrt image" option in the image position settings on the "Media Enrichment" tab.

"Template processing order" control in the feed settings, allowing users to specify the processing order for post title, content, and excerpt templates using a drag-and-drop interface.

"Convert Markdown to HTML" option which is applied to the AI-generated texts.

Shortcode processing algorithm.

The feed settings panel interface has changed. Some options have been moved from the "Basic" tab to the "Advanced" tab, and the post template settings are now available from their "Post templates" own tab.

Amazon search depth is now set to 2 by default.

Full-Text RSS site configuration (https://github.com/fivefilters/ftr-site-config).

WordPress 6.4.3

Telegram import issue.

The [gpt_article] shortcode now supports the "mistralai/mixtral-8x7b-instruct" model.

Functionality of the plugin in "auto" mode when the WordPress scheduler is in use.

Fixed an issue where using the [gpt_article] shortcode resulted in the duplication of section headings within their text content.

The [or-text] shortcode to access all language models available through OpenRouter.ai.

The "RSS.app JSON" preset.

OpenAI GPT-4 Turbo, Google Gemini Pro and Anthropic Claude models are now accessible via OpenRouter.ai for the [gpt_article] shortcode.

A problem that caused a "[gpt_article] was unable to generate an {n}-section list" error.

Cumulative bug fix.

DALL-E 3 is now the default model for the [dalle] shortcut.

The [stable_diffusion] shortcode now uses the "enhance" preset by default.

The "heading" parameter in the [gpt_article] shortcode can now contain HTML attributes. For example: [gpt_article topic="%post_title%" heading="div class='heading' id='article'"].

WordPress 6.4.3.

Google Gemini support via [gemini] and [gpt_article] shortcodes.

"Create from the image media attachment" now uses media content instead of media thumbnail to generate post thumbnails.

If a translating service is enabled, the meta description is now translated as well.

Anthropic Claude support via [claude] and [gpt_article] shortcodes.

The "structure" parameter in [gpt-article] WordPress-style [gpt-article] shortcode.

Shortcodes checking against errors caused by internal double quotes.

Wrong default image name generated by [dalle], [stable_diffusion] and [pixabay] shortcodes.

The Amazon search depth limit on the Accounts page. Use lower values to import only the latest products and reduce the risk of your host IP being blocked, resulting in a 503 error.

"Avoid AI detection" mode in WordAi content spinner.

The processing of the feed is stopped when a GTP error occurs.

1344x768 is now the default image size for the [stable_diffusion] shortcode.

The following older GPT models have been removed because they are no longer supported by OpenAI: ada, babbage, curie, davinci.

[gpt-article] shortcode.

Comulative bug fixes and code optimizations.

Post template placeholders are now allowed in the "Post tags" options field.

HTML tags balancing is now performed after full-text article extraction.

If the newly created post does not have any images, the post thumbnail generation will now use the images from the original source article.

Fully compatible with WordPress 6.4.2.

%random_file[]% placeholder. It's useful for inserting random images from the specified directory into posts.

The WordPress root directory is now displayed on the General Settings page.

Streamlining of text-only WordPress post excerpts by removing redundant paragraph tags.

"alt" attribute added to images generated via Pixabay, Stable Diffusion and Dall-E image shortcodes.

Shortcodes are now pre-verified for errors caused by internal double quotation marks.

List of sites registered under CyberSEO Pro license displayed on "General Settings" page.

The plugin is now able to import archived CSV files specified in the "CSV or Text Dump" import menu. The zip, gzip and bzip formats are supported.

The "Skip header" checkbox in the "CSV or Text Dump" import menu. When checked, the first line of a CSV dump is considered a header and is skipped.

Negative prompts for Stable Diffusion.

GPT-4 Turbo support in the AI Article Generation, Auto-Comments and CyberSEO GPT Spinner menus.

Amazon product duplicate checking method.

Amazon images are now imported at maximum size.

Post content is now automatically stripped of structural HTML elements (doctype, html, head, body tags).

Spintax parsing in templates (title, post, excerpt, meta description, header, footer).

Formatting of code blocks in imported posts.

"Newlines to paragraphs" settings.

OpenAI DALL∙E 3 text-to-image model.

If the OpenAI GPT shortcode fails, the post will not be added.

Stable Diffusion 1.0 XL is default model now.

Fully compatible with WordPress 6.4.

Shortcodes with additional arguments are now supported in post templates.

Support for using placeholders in post templates in custom taxonomies.

Media library image cleanup issue.

The "Republish existing posts" option now preserves the original post status.

The operation of the built-in synonymizer has been greatly improved, adding the ability to escape delimiters ("|"), etc.

Improved post title generation algorithm when importing RSS feeds.

A bug where the plugin did not properly handle comma as a decimal separator for EUR prices on Amazon, now correctly converting it to a period for accurate pricing.

WordPress Media Library image handling.

The accuracy of the %gpt_keyphrase% shortcode.

Unified shortcode names.

The "YouTube video" menu to the "Media Enrichment" tab.

OpenAI GPT 3.5 Turbo Instruct model.

OpenAI GPT 3.5 Turbo Instruct is now the default model for the GPT Content Spinner.

The "Shorten excerpts" settings are now applied prior to translation.

Added GPT-4 model support for the OpenAI GPT spinner.

"Google image" menu to the "Image generation" tab.

%gpt_keywords%, %gpt_keywords_en%, %gpt_keyphrase%, %gpt_keyphrase_en% placeholders.

HTML template parser has been optimized.

"Save to custom field 'thumb'" option on the "Image generation" tab.

OpenAI GPT spinner.

Nested placeholder handling in shortcodes.

Post thumbnail generation from "thumb" custom field when "Add to the Media Library" enabled.

Textorobot spinner.

Full-text article extraction algorithm using container tags.

Custom field generation from post content using outer element and regex.

Security fix: added protection against CSRF attacks during the "Export selected feeds" operation.

By default, content filtering is now applied after full-text article extraction.

Additional filtering logging.

All accumulated bugs related to content filtering.

The "republish existing posts" routine.

Issue with escaping Stable Diffusion JSON prompts.

Bug in the automatic generation of comments with OpenAI GPT.

Insertion of images found with AI into articles.

Amazon product search scrapping algorithm.

The "Image generation" settings are now saved in presets.

API function cseo_get_random_image and placeholder %random_image[]% now accept a list of keywords separated by commas.

The accuracy of image matching with AI.

user_agent enforcement when downloading images from remote hosts.

A problem with punctuation symbols in the content shortening routine.

Bug causing missed img class for images created in the "Image generation" menu.

Support for Yandex Translate API v2.

Conversion of absolute URLs to relative fragment identifiers in imported HTML documents was implemented.

Support for national characters in UTF-8 encoding.

Minor bugs.

Post template handling.

Stable Diffusion XL v1.0 engine.

Bug in "Remove outer HTML elements"

Saving images locally for relative URLs.

"Shuffle update times" button.

The %post_content_notags%, %post_content_notags[]% and %post_excerpt_notags[]% shortcodes.

NewsAPI preset.

New Stable Diffusion engines and options.

Amazon product import routine.

The OpenAI GPT spinner now uses the GPT-3.5 Turbo 16K model.

Implemented post meta description generation for the supported SEO plugins: Yoast SEO, Rank Math SEO, The SEO Framework, SEOPress.

The %post_tags% and tegories% shortcodes.

Processing of Google News feeds.

The "Remove outer HTML elements" preference.

The "Skip the 2nd check" option.

CyberSEO Pro's shortcodes can now also be used in the "Post title" and "Post excerpt" templates.

The "Hold for approval", "Max comments" and "Must be newer than ..." options on the "Auto-comments" options page.

New %post_date% and %post_date[]% placeholders.

The "Image generation" tab has been redesigned.

Optimized the shortcode processor.

Pixabay image search now accepts comma separated keywords.

Added the "Use alternative image source" option to the "Post thumbnails" menu.

The "Preferred GPT image source" menu is now available when "Find a relevant image using GPT" is selected.

Added Spintax support for post tags.

OpenAI GPT Tubro 3.5 16K model support.

Bug in the post template placeholder parser.

“Find a relevant image using GPT" post thumbnail option.

Filtering by post link.

Error in the "YouTube Atom feed" preset.

CyberSEO Pro GPT content spinner.

Content filtering order now can be adjusted.

Workflow order has been reorganized.

Adjustable delay between adding new posts.

Content Syndicator.

Pixabay settings.

Missing preset user agent in feed preview.

Auto tags

DALL-E and Stable Difussion images are now generated after post content filters are applied.

Duplication check rules have been tightened.

The "Newlines to paragraphs" option.

The third-party spinner options have been moved to the feed settings page.

Optimized the OpenAI GPT spinner.

All 3rd-party spinner credentials moved to the "Accounts" page.

The "Don't preprocess on preview" option on the "General Settings" page.

DeepL, Google Translate and Yandex Translate API key imput fields moved to the "Accounts" page.

Post template parser.

Import of extrenal CSV files.

Tags to WooCommerce option.

Categories to WooCommerce option.

AliExpress CSV preset.

The OpenAI GPT content spinner.

The "Use custom settings option" in the "Extract full text articles" drop-down menu.

The "HTML tags to strip" option.

Extended %post_content[max_words]% and %post_excerpt[max_words]% placeholders.

"Link handling", "Sanitize content", "Balance HTML tags" and "Shorten post excerpts" moved to "Advanced" tab.

Parse WordPress archives.

WordPress KSES post filtering issue.

Post tags and publication dates are now filtered before 3rd party service requests.

Issue with block delimiters in OpenAI article assignments.

The cyberseo_openai_shortcode() function is now availabe in "Modifycation Tools".

Shortcode parser.

WordPress 6.2.

Spinax is now supported in fixed string values for post custom fields.

Processing order for GPT assignments.

Duplication control routine.

Some default settings were changed.

Minor bugs.

It's now possible to set the API request limit for an hour, a day, and a month.

Hourly limits for API requests on the Accounts settings page.

OpenAI GPT-4 model support.

%pixabay[]% placeholder for post templates and alternative thumbnail image source.

Minor bugs.

Media library cleanup routine.

"Add to the Media Library" option.

Parser for Amazon, Tiktok, Twitter, Facebook, Instagram, Telegram.

OpenAI GPT-3.5 Turbo model is now available.

OpenAI GPT is now supported by the auto-comments generator.

Spintax is now supported by the auto-comments generator.

Amazon product import routine.

WordAI settings option to avoid AI detection.

Post template placeholder.

The SpinnerChief settings panel and the processing code rewritten in accordance with the new API.

A link to source option functionality.

A bug in CSV and raw text dump import.

YouTube thumbnail selection, when no pictures found for featured image.

Added API key check to skip image generation for DALL∙E 2, Stable Diffusion and Pixabay, if it's not set.

Shortcodes.

The parser code has been sufficiently optimized.

Bug in the XML feed editor related to incorrect error message output.

Google News RSS feed parser. The link redirection is not a problem anymore.