> ## Documentation Index
> Fetch the complete documentation index at: https://slatehq.com/docs/llms.txt
> Use this file to discover all available pages before exploring further.

# Bot Analytics

> See which AI bots read your website, which pages they open to answer people, what they can't reach, and which pages AI reads but never cites.

Bot Analytics shows every visit AI bots make to your website: ChatGPT opening a page to answer someone, search bots building the index that AI answers cite, and training crawlers. See which pages they read, which they miss, and what blocks them.

> [!NOTE]
>
> Bot Analytics is available to selected workspaces. Ask your Slate contact to turn it on for your workspace.

## Why It Matters

An AI answer can only cite a page that an AI bot has read. Bot Analytics tells you whether the bots reach your important pages, how fast they find new ones, and whether a firewall rule or an error keeps them out.

## Where to Find It

Open **Insights → AI Visibility** in the Slate sidebar, then select **Bot Analytics** under **Website**.

Bot Analytics counts the visits Cloudflare sends to Slate. Until an admin connects Cloudflare, the page shows **See which AI bots read your site**. See [Set up Bot Analytics](https://slatehq.com/docs/bot-analytics/connect-cloudflare.md).

## Bot Types

Slate recognizes AI bots by name and sorts each one into a type:

| Type | What the bot is doing |
| - | - |
| **Answering people** | An assistant like ChatGPT opened the page to answer someone's question right then. |
| **AI search** | Building the index AI answers cite from. |
| **AI training** | Collecting text to train future models. |
| **AI agent** | An agent browsing for a person. |
| **Search engine** | Googlebot and bingbot, which also feed AI Overviews and Copilot. |
| **Other bots** | Other crawlers, such as SEO tools. |

Slate also checks each request against the IP addresses the bot's company publishes. A request that claims to be a bot but fails the check is spoofed, and Slate leaves it out of every number.

## Filters

* **AI company**: one company, several, or **All AI companies**.
* The date pill: **Today**, **Yesterday**, **Last 7 days**, **Last 30 days** (default), **Last 90 days**, **Since connected**, or **Custom range…**. Turn on **Compare to previous period** to see changes. Days are in UTC.
* **Filters**: narrow by **Bot type**, **Bot**, **Hostname**, **Folder**, **Status code**, or **Verification**. **Verification** is **Exclude spoofed** by default. Choose **Verified only** or **Spoofed only** to change it.

Click **Reset** to clear the filters. The page has five views: **Overview**, **Bots**, **Page coverage**, **New pages**, and **Access & errors**. Your filters stay when you switch views.

## Overview

A sentence at the top sums up the range, for example how many times AI assistants opened your pages to answer people, and which company and page lead.

| Tile | What it shows |
| - | - |
| **AI visits** | Every request from an AI bot: training, search, agents, and answering people. |
| **Visits to answer people** | Requests where an assistant opened a page to answer someone. |
| **Pages read by AI** | Different pages at least one AI bot read in the range. |
| **Time to first AI crawl** | Median time from publishing a page to its first AI search or training visit, for pages published in the range. |
| **Errors to AI bots** | Share of AI bot visits to your pages that got an error. Scanners asking for files like `/.env` are left out. |

The chart shows a metric split by **Bot type**, **AI company**, or **Bot**. Click the chart title to change either. Markers show when a page was published or updated in Slate.

**AI companies** lists each company's visits, visits to answer people, pages, and errors. **Pages opened to answer people** lists the pages assistants opened most, and whether AI answers to your tracked prompts cite them (**Cited** or **Not cited**).

## Bots

Each bot shows its **AI visits**, **Pages**, **Last seen**, a **Trend** line, what your **robots.txt** says (**Allowed**, **Blocked**, **No robots.txt**, or **Not checked yet**), and its **Errors**. Click a bot to see the pages it reads most and the status codes it got back.

Switch to **Request log** to see each bot request, newest first. Slate keeps every request for 30 days and daily totals for 13 months. It checks IP addresses and then drops them, so the log never shows them.

## Page Coverage

Page coverage answers one question at a time. Pick it from the menu at the top left:

* **Crawled, not cited** (default): read by AI, but no tracked prompt cites it.
* **Never read by AI**: no AI bot read it in the date range.
* **Cited, not re-read in 30 days**: answers may quote an old version.
* **Errors to AI bots**: an AI bot got an error in the last 7 days.
* **Most opened to answer people** and **All pages**.

To keep a question with your search, click **Save current view…**. Saved views are shared with everyone in the workspace. Click a page for its visits, checks, and where it's cited.

## New Pages

New pages lists pages published or updated in the date range, with the time from publishing to the first visit of each bot type, and when a tracked prompt first cited the page. It uses your Slate [Pages](https://slatehq.com/docs/pages/overview.md) list.

## Access & Errors

* **Problems** lists what keeps AI bots out, such as a Cloudflare rule that blocks or challenges them, old URLs that return a 404, or pages that fail with a server error. Each problem says why it matters and how to fix it. For a Cloudflare rule, click **Copy the fix**, then **Open the rule in Cloudflare ↗**.
* **Who can get in** shows, for each bot, what robots.txt says, what Cloudflare does, and how many of its requests worked.
* **Errors** groups errors into **Your pages**, **Other page URLs**, and **Files and scanners**. It also shows your robots.txt history and whether your site has an llms.txt file.

## Share and Download

Click the share icon next to the page title:

* **Copy link to this view** copies a link that opens the same view.
* **Download CSV** downloads every bot request in your filters.
* **Ask Cowork for a summary** opens Cowork next to the page.

## Status

The status next to the page title shows the Cloudflare connection, for example **Cloudflare · 5 min ago**. Click it to see the details. If Cloudflare stops sending data, a red **Nothing from Cloudflare** chip appears. Hover over it to see why.

## Tips

* Check **Crawled, not cited** first. These pages reach AI already, so improving them can turn reads into citations.
* If **Errors to AI bots** rises, open **Access & errors → Problems** before you change content.

## What's Next

* [Set up Bot Analytics](https://slatehq.com/docs/bot-analytics/connect-cloudflare.md): connect Cloudflare
* [Citations](https://slatehq.com/docs/ai-tracker/citation-analysis.md): the pages AI answers cite
* [Pages](https://slatehq.com/docs/pages/overview.md): how each page performs in search and AI


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.