---
title: "Kitbase Data — What AI Engines Actually Cite"
description: "Aggregated statistics on what AI engines cite: which domains ChatGPT, Gemini, Google AI Overviews, Google AI Mode, Perplexity and Claude cite most, how often they cite anything, and how many sources an answer carries."
canonical: https://kitbase.dev/data
---

Kitbase Data

# What AI engines actually cite

**No engine has cleared publication yet.**

No months have cleared the publication thresholds yet. Reports appear here after the nightly aggregate first runs — a month is published only when it clears minimum-volume and diversity thresholds.

As of 28 Aug 2026, 04:20 UTC · [Every figure as JSON](https://api.kitbase.dev/public/insights/ai-citations) · How this is made

These pages are built from Kitbase AI Visibility tracking runs — real answers from real AI engines, most of them read from the same consumer interface a person sees. Everything is aggregated across every customer's tracked prompts and reported as percentages: nothing here identifies a customer, a project, a prompt, a brand or a person, and each customer's own domain is excluded from their own answers. Recomputed nightly, and a month is published only once it clears minimum-volume and diversity thresholds.

## How engines behave

[

### Answers with citations

How often an engine cites anything at all

Engine behaviour

](/data/answers-with-citations)[

### Average sources per answer

Sources attached to a citing answer

Engine behaviour

](/data/average-sources-per-answer)

## The same tracking, pointed at your brand

Which of your pages get cited, by which engine, against which competitors.

[Start free](https://app.kitbase.dev/signup) [How AI Visibility works](/ai-visibility)