Fixed-fee sprints · evidence included

Your product works.
AI agents can't find it.

We audit and fix the layer agents read to find, compare, and call products like yours. llms.txt, API catalogs, MCP server cards, markdown negotiation, robots policy.

Get a free readiness scan Send your domain. You get the 20-check evidence back, no call needed.

The numbers are brutal

Cloudflare scanned 200,000 domains for agent readiness. AI content signals showed up on about 4 percent. MCP server cards and API catalogs combined appeared on fewer than 15 sites.

~4%

of 200k domains pass AI content signal checks

<15

sites with MCP server cards or API catalogs

8 of 12

largest x402 sellers scored 7/20 or worse

We ran our own scanner across the twelve largest x402 payment sellers this month. The best passed 16 of 20 checks. Eight of the twelve scored 7 or fewer. These are companies whose customers are literally agents, and the agents cannot discover them through standard checks.

What I fix

Every sprint is scoped to what your scan shows. The usual work:

llms.txt + resource mapAn agent-facing summary that mirrors your real pages

API catalog/.well-known/api-catalog so catalog-driven agents find your endpoints

MCP server cardAt the standard location, if you expose tools

Link headersTying your discovery files together

robots policySays what agents may do instead of leaving them to guess

Markdown negotiationAgents get clean text on your key pages

Action safety copyWhich routes are read-only, which spend money, which need human approval

Before/after evidenceScans saved, so the work is verifiable

How it works

Free scan. Send your domain. We run it through 20 checks and send back the evidence. No call needed.
Audit. A ranked findings doc, plain English, with paste-ready fixes where possible.
Sprint. We implement the discovery layer and hand you the diff.
Proof. Before and after scan results, and the Cleared for Agents mark to display once you pass.

Rates

Fixed fee, scoped before work starts. No retainers unless you ask for monitoring.

$300

Audit

20-check scan, ranked findings, paste-ready fixes where possible.

$750+

Implementation sprint

The full discovery layer implemented, before/after evidence included.

Sites that pass display this mark:

Who runs this

Cleared for Agents is run by Hu White, who builds agent payment infrastructure and paid MCP tooling at agenttoll.dev. The audit runs on our own projects first: our marketplace scored 7 of 20 before we started fixing it. That score is the reason this service exists.