# Analytics

Every library has reading statistics, on every plan. Writers of the library find them under **Analytics** in the **Library** menu.

## What is counted

For each page view of a published page we record the page, the library, the language, the day and hour, the referring site, the country, the kind of device, and the source:

- **People on the web**: readers in a browser.
- **AI crawlers and Markdown readers**: the established AI crawlers, and requests for the Markdown version of a page or for `llms.txt`. They are counted on their own, so you see how much your documentation is read by AI.
- **MCP reads**: for the MCP server that is coming (see [AI assistants, API and MCP](https://docsahoy.docsahoy.eu/docs/ai-and-api)). It stays at zero until then.

Search engines and other bots are not counted.

## What is never recorded

No cookies, no IP address, no user agent, no identifier of any kind. The country is looked up from the IP address in memory during the request and the address is then discarded. Your readers have nothing to consent to, so your sites need no cookie banner.

## The dashboard

Choose the last 7, 30 or 90 days, or a range of your own. The figures are views, people on the web, AI crawlers and Markdown readers, and MCP reads; then the top pages, the referring sites, the countries and the devices. Referrers, countries and devices count people on the web only: crawlers have none of those that mean anything.

Days are UTC days, and the newest day shown is yesterday: today's views appear tomorrow.

## How long it is kept

Detailed page views are kept 90 days. Each day is also added up into daily totals, which hold no personal data and are kept as long as the library exists. The pricing page lists how much analytics history each plan includes.

## Searches

What readers type into a library's search box is recorded too, with the number of results. See [Search](https://docsahoy.docsahoy.eu/docs/searching).