Open data

Coming soon No dataset is published on this page yet. When it is, every file will state how and when it was collected.

What we plan to publish

• The question sets we test (Chinese and English), with the intent behind each question.
• Which engines were queried, how (API or web), how many runs, and on which date.
• Raw answers and cited source URLs, so anyone can re-check our numbers.
• Summary tables: mention rate and cited domains per question and engine.

Where

Data files will be published on this page and in a public GitHub repository alongside our open-source audit tool, citeshine-audit. Links will be added once they exist.

Our rules for data

• Proxy measurements (for example web search results used as a stand-in for an AI engine) are always labelled as proxies.
• Single-run answers are never presented as stable results.
• Numbers from vendors or media that we have not verified are marked as unverified.