Agent Readiness
Be navigable by agents.
The next wave of users isn't humans — it's agents acting on their behalf. Your homepage isn't built for them. We help you ship the machine-readable maps that are.
Why agent readiness matters
The next wave of traffic isn’t human.
An agent is asked to “book a table for two tomorrow night near London Bridge.” It hits ten restaurant sites. Nine of them are HTML soup it has to interpret by reading — slowly, with errors. One has /.well-known/agent-skills/index.json declaring a check-availability endpoint with parameters and an OAuth flow.
Guess which restaurant gets the booking.
This isn’t hypothetical. ChatGPT actions, Claude tool use, Anthropic’s Computer Use, and every emerging agent platform all converge on the same need: machine-readable manifests of what your site can do, where its endpoints live, and how to authenticate against them.
We test all of it — every well-known endpoint, every spec — and generate the missing files so you can ship them in an afternoon.
Agent Skills
A machine-readable index of what your site can do.
The /.well-known/agent-skills/index.json file is becoming the standard way for agents to discover capabilities. We auto-generate one from your existing API endpoints or sitemap, validate against the spec, and let you edit it before deploy.
- Auto-detection of public API endpoints
- Schema validation against the agent-skills spec
- Per-skill parameter documentation
- Versioning — every deploy is recorded and rollback-able
- WordPress plugin auto-deploys to /.well-known/ (Plus + Pro)
{
"version": "1.0",
"site": {
"name": "Acme Studios",
"url": "https://acme.studio"
},
"skills": [
{
"id": "check-availability",
"name": "Check booking availability",
"endpoint": "/api/availability",
"method": "GET",
"parameters": {
"date": "ISO 8601 date",
"service": "product | lifestyle"
}
}
]
}Model Context Protocol
Connect your tools to AI agents via MCP.
MCP — Anthropic’s open Model Context Protocol — is becoming the standard for connecting AI agents to external tools and data. We help you publish a discoverable server card so any MCP-aware client can find your server and connect.
- /.well-known/mcp/server-card.json generator
- Transport detection (SSE, HTTP, WebSocket)
- Auth metadata wiring (OAuth, API key, mTLS)
- Capabilities declaration (tools, resources, prompts)
- Validates against the public MCP spec
{
"version": "1.0",
"name": "Acme Studios MCP Server",
"endpoint": "https://mcp.acme.studio",
"transport": "sse",
"auth": {
"type": "oauth",
"metadataUrl":
"/.well-known/oauth-protected-resource"
},
"capabilities": {
"tools": true,
"resources": true,
"prompts": false
}
}Per-bot access rules
Decide who reads what, with surgical precision.
A blanket Disallow: / cuts off both AI training crawlers and answer-engine citations. You probably want different policies for each. We let you define per-bot rules with explicit allow lists, deny paths, and rate limits — then compile them into a clean robots.txt.
- Granular per-bot rules for 21 known AI crawlers
- Path-level allow/deny patterns
- Optional rate-limiting (Pro tier)
- Generated robots.txt is auditable and human-readable
- WordPress plugin auto-deploys (Plus + Pro)
User-agent: GPTBot Disallow: /pricing/internal Allow: / User-agent: ClaudeBot Allow: / User-agent: PerplexityBot Allow: / Crawl-delay: 5 User-agent: * Allow: / Sitemap: https://acme.studio/sitemap.xml
What gets scanned
The exact checks we run
Five dedicated agent-readiness checks plus three foundational ones — every scan runs all of them.
11 checks · 13.0 weighted points
robots.txt
Site has a valid robots.txt with explicit AI bot directives.
sitemap.xml
Sitemap is present, valid, and discoverable from robots.txt.
Link headers
Server returns Link headers pointing to canonical, sitemap, etc.
AI bot access rules
robots.txt explicitly addresses major AI bots (GPTBot, ClaudeBot, etc.).
Agent Skills index
A valid /.well-known/agent-skills/index.json is present.
MCP Server Card
A valid /.well-known/mcp/server-card.json is present.
API catalog (RFC 9727)
Site exposes an API catalog at /.well-known/api-catalog (RFC 9727).
MCP manifest (deep)
Validates /mcp.json or /.well-known/mcp.json — required fields, tool schema correctness, agent-callability.
OAuth discovery
OAuth metadata endpoints (RFC 8414, 9728) present where APIs require auth.
Agent skills (deep)
Validates /.well-known/agent-skills.json — each declared skill must be agent-callable with name, description, and a target.
API discoverability
AI agents can find your public API via OpenAPI link tags, HTTP Link headers, or well-known paths.
Free · No signup · 30 seconds
See your score now.
One scan tells you exactly what AI engines see when they look at your site, and the specific files we’ll generate to fix what’s missing.