# Why a curated agent-ready directory beats a raw crawl

Agents need **trustworthy** pointers to sites that actually expose machine-readable specs—not every URL that happens to rank on a keyword.

## The crawl problem

A fully automated directory can index thousands of domains quickly, but it also publishes:

- Stale or broken `llms.txt` files
- Marketing pages with no MCP or API surface
- Sites the owner never opted into

Humans and agents both waste time filtering noise.

## What we do instead

**AI Agent Pages** stays curated:

1. **Submit or discovery** — Founders submit; we also probe promising URLs in an admin-only discovery queue.
2. **Human review** — Pending listings are approved or rejected with notes.
3. **Claim** — Discovered listings stay **unclaimed** until someone signs in with a matching domain email and claims ownership.
4. **Reference listings** — Editorial examples are labeled clearly and are not membership claims.

That flow protects your brand and gives agents a higher signal catalog.

## SEO and education still matter

Curated does not mean secret. We publish [guides](/guides) on llms.txt, MCP server cards, and agent-ready SaaS setup so you can improve before you submit.

## List your site

When your specs are in place, [submit for review](/submit). If we discover you first, watch for outreach—or [claim](/badge) after approval.