In the wake of AI advances and the constant emergence of new, increasingly powerful language models, the CyberSEO Pro plugin continues to evolve – introducing new features and improving existing ones. For example, after the release of OpenAI’s GPT-4o mini-model, significant changes were made. Since you might not be able to discover all the new features without delving into the relevant settings sections, we’ll explore the most important changes below.
AI Translate
A powerful translation tool, AI Translate, has been added to both CyberSEO Pro and RSS Retriever plugins. Now, alongside traditional options like DeepL, Yandex and Google, users can opt for an AI model for translation. Currently, OpenAI’s GPT-4o mini is available, accessible via the official API or OpenRouter’s unified API. GPT-4o mini has several advantages:
- Extensive contextual coverage (128k input, 32k output)
- High-quality translation
- Support for numerous languages
- Cost-effectiveness compared to competitors
AI Spinner
The AI Spinner has received a substantial upgrade in CyberSEO Pro and made its debut in RSS Retriever. Powered by GPT-4o mini, this feature now supports multiple languages, including German, Spanish, French, Chinese, and Turkish. The spinner preserves the original HTML structure of the text, ensuring unparalleled quality in content transformation.
Auto-comments and AI Article Generation
CyberSEO Pro’s Auto-comments and AI article generation functions now support a wide array of language models, including:
- OpenAI GPT-4o mini (official API, OpenRouter)
- GPT-4o (official API, OpenRouter)
- OpenAI GPT-4 32k (official API, OpenRouter)
- GPT-3.5 Turbo 16k (official API, OpenRouter)
- Anthropic Claude 3.5 Sonnet (official API, OpenRouter)
- Google Gemini Pro 1.5 (official API, OpenRouter)
- Meta Llama 3.1 405B (OpenRouter)
- Mistral Large (OpenRouter)
Of particular note is Meta Llama 3.1 405B-the first open-source frontier-level model, released on July 23, 2024. It supports eight languages and has an extended context of up to 128K. Meta Llama 3.1 405B support is already implemented in the CyberSEO Pro, RSS Retriever and AI Autoblogger plugins.
%gpt_image_prompt% placeholder
The new %gpt_image_prompt% placeholder parses the article text and generates a prompt to create an appropriate image. It’s compatible with graphics models such as Midjourney, DALL-E 3, and Stable Diffusion. This placeholder uses the default AI model selected on the General Settings tab.
Optimal use of %gpt_image_prompt% is within [midjourney], [dalle], and [stable_diffusion] shortcodes. For example:
[midjourney prompt="%gpt_image_prompt%" aspect_ratio="16:9"]
%gpt_image_prompt% can also be used in prompt fields on the Media enrichment tab. In this case the placeholder generates a unique prompt for each section of text before which an image is inserted. This is a very powerful tool to graphically illustrate text imported from third party sources or generated from scratch using AI.
Thanks to the use of models with large contextual windows and accessible APIs, such as OpenAI GPT-4o mini, Google Gemini Flash 1.5, and Anthropic Claude 3 Haiku, the cost of using this feature remains minimal.