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

# RankRaftBot

> What RankRaftBot is, what it requests, and how to allow or block it.

You're probably here because you saw us in your access log. Here is exactly what we are and what we asked for.

RankRaft AI is an AI visibility tool. Site owners use it to find out whether AI answer engines can read their site, and whether those engines recommend them. We only fetch a site at the request of someone who has added that domain to their own RankRaft AI account.

## What you will see in your logs

We send three shapes of user agent, all linking back to [rankraft.ai/bot](https://rankraft.ai/bot):

| User agent | Purpose |
| - | - |
| `RankRaftBot/1.0 (+https://rankraft.ai/bot)` | Our own identity, used for ordinary fetches such as your homepage and `robots.txt`. |
| `RankRaftProbe-Unknown/1.0 (+https://rankraft.ai/bot)` | A crawler nobody has written a rule for. It tells us whether a block targets AI crawlers specifically or all automated traffic. |
| `<BotToken>/1.0 (+https://rankraft.ai/bot) RankRaft-AccessCheck` | The [AI access check](/agent-analytics/bots/ai-access-check). |

### The access check is the one that looks alarming

To tell you whether **you** are blocking GPTBot, we have to ask your server what it would do if GPTBot knocked. So that third user agent begins with the token of the bot being tested. A line like `GPTBot/1.0 (+https://rankraft.ai/bot) RankRaft-AccessCheck` in your log is **us**, not OpenAI.

This is deliberate and we don't hide it. Edge rules match on the bot token as a substring, so the prefix is what makes the test meaningful, while the suffix and the URL make it obvious who actually called.

Other RankRaft user agents you may see:

* `RankRaft-Test/1.0`: test events and the cache check from the WordPress plugin, on your own server.
* `RankRaft-Callback/1.0`: a signed request telling your WordPress plugin to resume logging.

## What we request

* Your `/robots.txt`, your homepage and a small number of requests to a single page.
* Requests are spaced **400 ms** apart. We don't crawl your whole site, and we don't follow links.
* We send no cookies, submit no forms and run no JavaScript. Every request is a plain GET or HEAD.
* A check runs when a customer asks for one, during setup, and on a monthly schedule. Not continuously.

## How to allow us

Nothing is required. If you want to be explicit, `RankRaftBot` is the token to allow. Our access check deliberately identifies as the bot under test, so a rule written against `GPTBot` will also stop the check. That's the correct result, because it's what we're measuring.

## How to block us

Add this to your `robots.txt` and we'll stop:

```txt robots.txt theme={"system"}
User-agent: RankRaftBot
Disallow: /
```

We honour it. If you believe we behaved badly, or want us to stop immediately rather than at the next check, write to [abuse@rankraft.ai](mailto:abuse@rankraft.ai) with a log excerpt.


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