> ## Documentation Index
> Fetch the complete documentation index at: https://docs.thepublive.com/llms.txt
> Use this file to discover all available pages before exploring further.
# Overview
> How AXP Edge serves AI crawlers a clean, agent-first version of your existing site.
AXP Edge is an edge-rendering layer that sits in front of your existing site on your CDN. It detects who's making a request and routes AI agents to a clean, structured, agent-first version of the page, while humans and search engines continue to see your site exactly as it is today.
## How it works
Requests are split by requester type at the edge, before they reach your origin:
* **Humans and SEO bots** pass straight through to your site, byte-identical to what they get today. Nothing about your site, SEO, or user experience changes.
* **AI agents** (ChatGPT, Gemini, Perplexity, and similar) receive a JS-rendered, token-optimized version of the same page: no navigation chrome or scripts, just the structured content an AI agent needs to read and cite.
## Not cloaking
AXP Edge only cleans up and restructures content that's already on the page. It doesn't show AI agents anything different in substance from what a human sees. Cloaking, by contrast, serves search engines different content to manipulate rankings. AXP Edge serves AI agents the same facts, in a form they can actually parse.
Search engines like Google and Bing are treated the same as human visitors. They receive your normal pages, so your existing SEO is unaffected.
## What changes for AI agents
| Signal | Effect |
| -------------- | -------------------------------------------------------------------------------------------------- |
| Tokens to read | \~90% fewer tokens than the human page, since navigation, scripts, and chrome are stripped |
| Rendering | Server-side rendered, so AI crawlers that don't execute JavaScript can still read the full content |
| Structure | Clean, semantic HTML built for citation rather than layout |
Brands running AXP Edge on already-indexed pages have seen roughly 15-20% more AI citations from the same underlying content, simply because it becomes easier for an AI agent to parse and trust.
## Enterprise-safe by design
* **Nothing changes for customers**: only AI tools receive the cleaned-up version.
* **It stays on your own domain**: there's no second page to manage, and citations point back to your real URL.
* **It fails safe**: if the edge layer is ever unavailable, requests fall straight back to your normal origin.
## Next steps
Connect your CDN and go live in under two hours.