Explore the new features that have recently been implemented in RAT:
New Classifier: GEO Score
Alongside our existing SEO classifier, RAT now offers a classifier for generative engine optimization (GEO). It calculates a score indicating how strongly a web page or AI source is optimized for being picked up and cited by generative AI systems. The GEO Score can be applied to organic results as well as AI sources, making it easy to investigate which content gets selected by AI search features.
Custom LLM Content Analysis
With our new LLM infrastructure, you can classify web pages, AI sources and AI answers with large language models of your choice. Simply add prompt rules in the study setup, and RAT will process the cleaned visible text of the collected items through any OpenAI-compatible API endpoint. This works with cloud routers such as OpenRouter as well as with a local Ollama instance, so you can also run your analyses entirely on your own infrastructure. A setup guide for Ollama and cloud APIs is available directly in the interface.
New Scrapers: Google Images, Bing Images and YouTube
RAT can now collect results from Google Images, Bing Images and YouTube. Together with the new assessment options, this opens up new possibilities for studies on visual and video search.
Assessing Images, Videos and AI Sources
The assessment process in RAT has been extended beyond classic web pages. From now on, participants can also evaluate images, videos and the sources cited in AI answers. This allows you to design human assessment studies that reflect today’s search engines more realistically, for example by comparing how participants judge image results, YouTube videos or the sources that an AI overview relies on.
To discover all the various features of RAT, visit the RAT Knowledge Base!