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

# URL rules

> How Locations decide which pages qualify. Operators, negations, and the per-rule case-insensitive toggle.

A URL rule is one row in the rule builder. It picks a **field**, an
**operator**, and a **value**. Locations combine URL rules with AND;
audiences use URL fields the same way for URL-scoped rules.

## The fields

| Field              | What it looks at                                             |
| ------------------ | ------------------------------------------------------------ |
| **URL**            | The full URL including protocol, host, path, query, and hash |
| **URL with query** | Everything except the hash                                   |
| **Path**           | Just the path (`/products/hats`)                             |
| **Query string**   | The `?…` portion, without the leading question mark          |
| **Hostname**       | The host only (`shop.acme.com`)                              |

## The operators

The rich fields (URL, URL with query, Path) get the full menu:

* **equals** / **does not equal**
* **contains** / **does not contain**
* **starts with** / **does not start with**
* **ends with** / **does not end with**
* **matches regex** / **does not match regex**

Hostname and query string have a reduced set, equals, contains, starts
with, ends with, and their negations. Regex on hostname is not on the
menu because in practice it is more error-prone than helpful.

## Case sensitivity is per rule

Each URL-family rule has a **Case-insensitive** toggle.

* New rules default to **on** (case-insensitive).
* Existing rules keep their prior behavior, the toggle is exposed so
  you can flip them at will.

Why the default flipped: `/Pricing`, `/pricing`, and `/PRICING` are the
same page to a human. A rule with case-sensitivity on will silently
skip visitors who type or share the "wrong" case, and the failure is
invisible on the results page because it looks like "no traffic on this
URL yet".

## Regex, if you must

Regex rules are full JavaScript regex. Anchors count, `pricing$`
matches a URL ending in `pricing`, `^pricing` matches a URL starting
with `pricing`. Case-insensitivity via the toggle takes precedence over
inline `(?i)` in your pattern.

Regex is powerful and hard to read. Prefer starts with / ends with when
they fit; reach for regex only for truly non-obvious patterns.

## Negations do not "include the rest"

**does not contain "checkout"** does not mean "include every page
without checkout". It means "if the URL contains checkout, exclude it";
the rest of your rules still have to match.

Two Locations rules ANDed together:

```
Path starts with /products
Path does not contain /admin
```

matches `/products/hats` but skips `/products/hats/admin`. The
"admin" rule alone does not include anything on its own.

## Where these live

* Inline in the experiment editor under **Where it runs**.
* Saved in a Location under **Site → Locations**. See
  [Saved lists](/targeting/saved-lists) for when to reach for a saved
  Location.

## Not in the URL rule set

* **Geo**, country, region, city are Audience rules, not URL rules.
  See [Geo](/targeting/geo).
* **Device / browser / OS**, Audience rules.
* **Cookie or JS conditions**, Audience rules, under Custom.

That separation is deliberate. URL rules decide *which page*; audience
rules decide *which visitor*. Keeping them apart makes broken tests
easier to diagnose.