Skip to main content
Agent Analytics gives you visibility into AI bot activity on your website. Using first-party data from your server logs is the most direct signal of how AI bots access your content. Crawl Insights reveals AI bot activity directly from your server logs – showing which bots access which pages and folders, how often they visit, the status codes they receive, and how all of this links to your AI visibility.

Connecting your data

To get started, go to Crawl Insights and connect a data source. We currently support two options:
  • Cloudflare Workers: for continuous, real-time log ingestion if your site runs on Cloudflare.
  • CSV/CLF File upload: upload a server log file directly to analyze a specific time period.
CSV/CLF files must have a .csv, .clf, or .txt extension.
Once connected, Peec AI identifies AI bot hits in your logs, categorizes each bot, and populates the dashboard. You can manage or disconnect your data source at any time from Settings.

Dashboard Overview

Once the data is loaded, your dashboard will appear. Here, you can access key information and use different tools to analyze your data.
Crawl Insights Dashboard
In the dashboard, you will be able to see:
  • Filters: Date, Platform (e.g., OpenAI, Google), Bot type (Training, Search, User Query, Other), and Bot (e.g., GPTBot, ClaudeBot)
  • KPI Summary: A row of key metrics at the top gives you a snapshot of your filtered data based on total bot visits, active bots, and failure rate (e.g., proportion of requests that returned an error (4xx or 5xx))
  • Crawl activity over time: The line chart shows AI bot visits, with a separate line for each AI model. Switch between hourly (only visible in a 3-day view), daily, weekly, and monthly views - daily is the default. Use this to spot trends, sudden spikes, or unexplained drops in bot activity.
  • Content section: Further down, you’ll see AI bot traffic broken down by folder and URL, showing which parts of your site AI bots visit most.

Crawl breakdown

The crawl breakdown will show you a set of bar charts that break down bot visits by different dimensions.
Crawl Breakdown
These dimensions are:
  • Platform: the AI vendor (e.g., OpenAI, Anthropic, Google)
  • Bot: the specific user agent (e.g., GPTBot, ClaudeBot, PerplexityBot)
  • Bot type: Training / Search / User query / Other
The Bot type indicates the purpose of the crawl:
  • Training bots collect data to build or refine AI models.
  • Search bots browse the web to find up-to-date information when needed.
  • User Query bots access your website to fetch content on behalf of a user.

Insight Details

Insight Details provides a URL-level breakdown of AI bot activity and source performance. You can access it from the in-page navigation bar or by clicking the See Details button above the By URL chart. In here you will see:
  • Filters: The same global filters apply (date range, platform, bot type, bot). Within the table, you can also filter by folder (drill into a specific section of your URL structure) and status code (show only URLs that returned a specific HTTP response (e.g., 200, 404, 500)). You can also use the table’s search bar to quickly find a specific URL.
  • KPI summary:
    • Total bot visits: Total requests across all URLs in the selected filters
    • Active bots: Number of distinct bots active in this view
    • Failure rate: Proportion of requests that returned an error (4xx or 5xx status code)
    • Top-visited folder: The top-level section of your site (e.g., /blog/, /products/) receiving the most bot visits
    • Top-visited URL: The single URL receiving the most bot visits

Visited URLs table

This table shows the pages on your site that can be crawled and actively indexed, with detailed information for each. This will give you a quick overview of which pages are receiving bot visits, how often they occur, and how this differs across all bots currently visiting that page. You will see:
  • URL: The requested page path from your domain
  • Folder: The primary section or directory the URL belongs to
  • Bot visits: Total AI bot visits to this URL in the selected period
  • Platforms: Bot visits by the vendor behind the bots requesting the URL
  • Status codes: The HTTP response the server returned to bots
Additionally, based on your prompt data, you will also be able to see:
  • Retrievals: The sum total of how many times this URL appeared as a source in AI chat responses
  • Citation rate: Average number of inline citations per chat when this URL is retrieved as a source
  • Topics: How many of your tracked topics had prompts where this page appeared as a source
Prompt data comes from the prompts you are tracking. It is not influenced by AI bot activity.

Status Codes

Status codes show how your server responded to AI bot visits, based on your connected log data. The overview table always shows a baseline set of status codes, even if they have zero visits in your log data. Additional codes appear only when present in your logs.
Status codeDescription
200OK
301Moved Permanently
302Found (Temporary Redirect)
304Not Modified
307Temporary Redirect
308Permanent Redirect
401Unauthorized
403Forbidden
404Not Found
410Gone
429Too Many Requests
500Internal Server Error
502Bad Gateway
503Service unavailable
504Gateway Timeout

FAQs

How do I connect?

Go to Crawl Insights and follow the setup prompt. The method depends on your hosting setup:
  • If you use Cloudflare, you’ll connect via a Cloudflare Workers, which streams your logs to Peec in real time.
  • If you use another provider or prefer to upload a specific dataset, use the CSV or CLF file upload option — just drag and drop your log file or select it from your file picker.

How do I disconnect?

Go to Crawl Insights → Settings. You’ll find the disconnect option in the danger zone. Disconnecting stops log syncing. Disconnecting does not automatically delete historical data.

What plans is agent analytics available on?

Agent Analytics is currently available to all accounts. Usage limits may be introduced in the future as we continue to scale the feature.

Does my Cloudflare plan affect how much data is collected?

Yes. If you’re on Cloudflare’s free plan, your Worker is limited to 100,000 requests per day. On high-traffic sites, this may mean not all bot requests are captured. Upgrading to a paid Cloudflare Workers plan removes this limit.

Can I request my data to be deleted?

You can request deletion of your data at any point by contacting our support team or by navigating into settings under the danger zone.

Privacy policy

To find more about our privacy policy, go here: Privacy Policy for Peec AI