Documentation
Integrate LinkAudit's AI-visibility scanning into your own tools and workflows.
What LinkAudit measures
LinkAudit fetches a page the way an AI crawler does — a plain HTTP request with no JavaScript executed — and reports whether AI engines can reach it, understand it, and quote it. That is a different question from search ranking, and a site can do well at one while failing the other.
The distinction matters because the crawlers that decide citation do not run JavaScript. GPTBot, ClaudeBot and PerplexityBot download script files and never execute them, so a client-rendered application serves them an empty shell. Everything below works against the HTML your server actually returns.
Integrations
The MCP server is the fastest way in if you already work in an AI editor: install it once and the scan becomes a tool your assistant can call while it is editing the site. The same package doubles as a CI linter, so a build can fail on AI-search visibility the way it fails on tests.
The files a scan generates
Every scan produces artifacts rather than only a score, each built from what is actually on the page:
| File | What it does |
|---|---|
| robots.txt | Explicitly admits the search-index crawlers that decide citation — OAI-SearchBot, Claude-SearchBot, PerplexityBot — not just the training ones. |
| schema.json | Organization and FAQPage JSON-LD so engines can identify you and lift your answers directly. |
| sitemap.xml | A current URL set for crawlers that use one to discover pages. |
| llms.txt | A plain-language summary of what your site is, for AI systems that read it. |
Where to start
If you want a result before reading anything, run a free scan — no signup, and the ranked fixes explain themselves. If you would rather wire it into a workflow first, start with the MCP server. For the concepts behind the checks, the guides cover them by industry in plain language.