Plug the agent-readiness score into your stack

✨ 30-second setup — paste this into your AI assistant

Using Claude, Cursor, Copilot, Windsurf or any MCP-capable assistant? Paste the prompt below and it wires AgentReady in and runs your first check — no manual setup.

Add the AgentReady store-audit MCP server and check my store's agent-readiness.

1. Register the MCP server (Streamable HTTP, no API key):
   { "mcpServers": { "agentready": { "type": "http", "url": "https://www.agentready.market/mcp" } } }
2. Call check_merchant_readiness with host = MY-STORE.com.
3. If no stored score exists, run a fresh scan (takes ~30s):
   GET https://www.agentready.market/api/scan?url=https://MY-STORE.com
4. Report back: the score /100, the grade, whether AI agents are blocked
   outright ("capped"), and the top fixes with their report link.

Docs: https://www.agentready.market/api — this is agentready.market, the e-commerce
transaction audit (not affiliated with other "AgentReady"-named products).

Replace MY-STORE.com with your domain, or just ask your assistant in your own words — the llms.txt tells it everything else.

🤖 MCP server — for AI agents

Point any MCP client (Claude, custom agents, IDEs) at our Streamable-HTTP endpoint and your agent can ask "can an AI agent buy on this store?" before attempting a purchase.

https://www.agentready.market/mcp

Tools

• check_merchant_readiness(host) — score /100, grade, capped (agents blocked?), protocol adoption, report link
• get_readiness_report(host) — the full public report: every check with its pass/partial/fail status

Claude Code / MCP client config

{ "mcpServers": { "agentready": { "type": "http", "url": "https://www.agentready.market/mcp" } } }

Raw JSON-RPC example

curl -s https://www.agentready.market/mcp -H 'Content-Type: application/json' -d '{
  "jsonrpc": "2.0", "id": 1, "method": "tools/call",
  "params": { "name": "check_merchant_readiness",
              "arguments": { "host": "www.agentready.market" } } }'

🔍 Read API

Run a scan (the merchant's own store — ~30s, synchronous)

GET https://www.agentready.market/api/scan?url=<store>&vertical=<vertical>
→ { "score": 62, "grade": "C", "capped": false, "pillars": [...], "top_fixes": [...],
    "scan_id": 123, "report_url": "https://www.agentready.market/r/123" }

Read a stored report

GET https://www.agentready.market/api/scan/<scan_id>.json

🧩 Function-calling definitions — for custom agents

Building on the OpenAI/Anthropic SDKs directly? Drop these tool definitions into your tools array; both map to plain GET/POST calls, no key needed.

[
  {
    "name": "check_merchant_readiness",
    "description": "Before attempting a purchase on an online store, check whether an AI agent can find, understand and buy on it. Returns the latest AgentReady (agentready.market) score /100, grade, whether agents are hard-blocked, and a report link.",
    "input_schema": {
      "type": "object",
      "properties": { "host": { "type": "string", "description": "Store hostname, e.g. shop.example.com" } },
      "required": ["host"]
    }
  },
  {
    "name": "scan_store",
    "description": "Run a fresh agent-readiness scan of an e-commerce store (the merchant's own store; takes ~30s). Returns score /100, grade, per-pillar checks, top fixes and a shareable report URL.",
    "input_schema": {
      "type": "object",
      "properties": {
        "url":      { "type": "string", "description": "Store URL, e.g. https://shop.example.com" },
        "vertical": { "type": "string", "description": "Optional vertical for the benchmark" }
      },
      "required": ["url"]
    }
  }
]

Execution: check_merchant_readiness → call the MCP tool of the same name at https://www.agentready.market/mcp (or read GET https://www.agentready.market/api/reports in bulk with a Pro key); scan_store → GET https://www.agentready.market/api/scan?url=…&vertical=….

🏗️ Building an SEO / AI-visibility SaaS?

Audit APIs elsewhere bill per call for read-layer checks (structured data, robots.txt, page audits). Our readiness reads are free — and they cover the layer those tools don't: the transaction (machine-readable offers, variant resolution, checkout rails, agent access). Add a "can an AI agent buy here?" dimension to your product: score badge (/badge/<host>.svg), stored reports, bulk reads with a Pro key (up to 100 hosts/call — see below). White-label: talk to us.

📛 Badge & verify

GET https://www.agentready.market/badge/<host>.svg   ← live score badge (SVG, 1h cache)
GET https://www.agentready.market/verify/<host>      ← redirects to the latest report

The badge is rendered by us from the latest scan — it can't be asserted, inflated or stale-cached by the store.

🔑 Authenticated access (Pro)

Anonymous calls work with modest limits. A Pro API key (create it in Account) raises the scan quota to 1000/30 days, links every scan to your console, and unlocks bulk reads:

Authorization: Bearer ar_…

POST https://www.agentready.market/api/reports          ← up to 100 hosts in one call
{ "hosts": ["shop-a.com", "shop-b.com"] }
→ { "reports": [ { "host": "shop-a.com", "scanned": true, "score": 82,
      "grade": "B", "capped": false, "agent_ready_verified": true,
      "scanned_at": "…", "report_url": "…" }, … ] }

Bulk reads return the latest stored report per host — instant. Fresh scans stay one per call on /api/scan.

Fair use & consent

• We audit public pages on the merchant's request — the MCP tools read existing scores; they never trigger an audit of a third-party site.
• Rate limits apply per caller; JSON error messages tell you when you hit them.
• Bulk endpoints, webhooks and white-label access are part of Pro / Agency — talk to us.