# Bot traffic policy (https://www.getfounderhq.com/docs/analytics/concepts/bot-traffic-policy)

Crawlers hit your site all day: search engines, AI crawlers, social
preview fetchers, SEO tools, uptime monitors, and your own scripts.
FounderHQ separates them from people before any number is drawn. You do
not configure this.

## How a visit is classified

Classification happens on the server, when the event arrives — not in
the browser, where anyone could switch it off. It uses user-agent rules
covering AI crawlers, search engines, social preview fetchers, SEO
tools, monitoring services, plain HTTP clients, and browser automation.
The web SDK adds one signal of its own: whether the page is being
driven by automation.

## What happens to a bot visit

Bot events are recorded, not thrown away. Each one is stored with a bot
classification and a synthetic guest identity that is shared by all
bots on that key. Two things follow:

- **No bot ever becomes a contact.** A crawler cannot create people in
  your audience, cannot enter a segment, and cannot enter a sequence.
- **Bot events do not count toward your plan's event limit.**

Bot traffic is also excluded from sessions, contact activity, and every
human metric in the product. Visitors, conversions, and revenue read
the same whether a crawler visited once or a million times.

## Why keep them at all

Because "which AI crawlers read my launch post" is a real question, and
2026 is a good year to be able to answer it. The events stay queryable
by classification, so AI and search crawler visibility is available
without ever touching your product metrics.

## The one limit

The bot signals travel on your publishable key, which anyone can read
from your page. So bot storage has its own capped pool, sized to your
plan's event limit and counted separately from your real events. Once
that pool is full, further bot events are refused at ingest and the
response says so explicitly. Your human events are never affected by a
flood of bot traffic.

## What this means for you

- Your first-week numbers are people, not preview fetchers.
- A viral link that draws scrapers does not inflate your bill.
- If a real browser is ever misread as a bot, tell support with the
  user agent. Classification is server-side, so a fix reaches everyone
  without an SDK release.

## Related

- [Sessions](/analytics/concepts/sessions)
- [Attribution](/analytics/concepts/attribution)
- [What FounderHQ analytics is](/analytics/getting-started/what-founderhq-analytics-is)