Six engines, measured separately
Averaging engines into one score hides the thing you need to know. A brand can be strong in two and absent from four, and those are different problems with different fixes. Each engine is reported on its own as well as combined, and each page states how that engine is collected.
Measure how often ChatGPT names your brand and which sources it cites instead. Collected through the official OpenAI API, never a scraped chat window.
Measure how often Claude recommends your brand and which sources it draws on. Collected through the official Anthropic API and live from day one on every plan.
Perplexity cites heavily, making it the clearest read on which sources decide your category. Collected through the official Perplexity API.
AI Overviews cite sources that are often not the pages ranking beneath them. Collected through licensed search data, as no public API exists.
Google AI Mode is Gemini-powered and retrieves from a different pool than AI Overviews. Tracked as its own engine, with its own weight.
DeepSeek answers collected from a United States hosted deployment of the open-weight model. No data is transmitted to the People's Republic of China.
Standalone Gemini is not listed because AI Overviews and AI Mode are both Gemini-powered, so tracking those two covers the Gemini family without a third overlapping subscription. Engine weights are published at /metric-docs.