---
title: Getting started
description: Create an account, add your site and read the first audit: which pages AI assistants can read and what to fix first.
id: getting-started
date: 2026-11-02
section: Basics
order: 10
---
Anole checks how ready your site is for AI assistants such as ChatGPT, Claude, Gemini and Perplexity, and helps you fix what holds it back. This guide takes you from a new account to your first audit.

## 1. Create an account

Go to the [sign-in page](/account.php) and enter your email. You'll get an email with a six-digit code and a link: type the code or click the link, whichever is easier. There's no password. Below the email field there are also buttons to sign in with Google and other accounts. Your account is created the first time you sign in.

## 2. Add your site

In the menu choose **Add site** and fill in:

- **Domain**, for example `example.com`, without `https://`.
- **Primary language** of your content.
- **Already have your own llms.txt? (optional)**: if your site already serves one, paste its address. Anole will compare it with your sitemap. Leave it empty if you want Anole to build the file for you.

After saving, Anole starts the first crawl. It reads your sitemap, so make sure `https://example.com/sitemap.xml` (or the address in `robots.txt`) lists your pages.

## 3. Read the audit

When the crawl finishes, the site page lists every address from the sitemap with a score from 0 to 100. For each page Anole checks, among other things:

- whether the page has a Markdown version that assistants can read without the layout,
- the canonical address,
- structured data (JSON-LD),
- the title, H1 and meta description.

Start with the pages that matter most for your business and have the lowest score.

## 4. Next steps

- [Deploy llms.txt](/en/docs/llms-txt/) so assistants find a map of your site.
- [Connect an AI assistant](/en/docs/connect-ai-assistant/) and ask it about your weakest pages directly in the chat.

> Anole shows what assistants can read on your site. It does not promise positions or traffic: the models decide what they answer.
