For AI agents and developers

Plug CanAIReadMe into your agent.

One engine, three doors: an MCP server, a REST API and the website. Agents can check how well AI systems understand any business website, read structured business facts with their sources, and hand users the fix.

MCP server

Streamable HTTP, current MCP protocol plus 2025-era clients, no sign-in needed.

https://canaireadme.netlify.app/mcp

Tool	Cost	Side effects	Returns
audit_ai_visibility	Free	Saves the analysis	0-100 AI Visibility Score, how AI would describe the business, what it can't determine, top issues with evidence.
get_business_profile	Free	Saves the profile	Business Passport v1: what the business is, sells and charges, how to buy or contact it, with provenance on every field.
get_fix_package	$49 one time	Creates a quote	Machine-executable fix list, rewritten copy, JSON-LD, meta tags, llms.txt, robots.txt rules. Returns a checkout link the user approves; never charges by itself.

Connect it

Claude (web and desktop): Settings, Connectors, Add custom connector, paste the URL.
ChatGPT: add it as a connector in developer mode with the URL https://canaireadme.netlify.app/mcp/free (free tools only).
Claude Code:

claude mcp add --transport http canaireadme https://canaireadme.netlify.app/mcp

Cursor, VS Code, Windsurf and other clients:

{
  "mcpServers": {
    "canaireadme": { "type": "http", "url": "https://canaireadme.netlify.app/mcp" }
  }
}

https://canaireadme.netlify.app/mcp/free serves the two free tools with no commerce in any result, for directories and hosts that don't allow selling digital goods. https://canaireadme.netlify.app/api/mcp keeps working for earlier integrations.

REST

Same engine and the same JSON as the MCP tools.

POST https://canaireadme.netlify.app/api/v1/agent/audit          {"url":"acme.com"}
POST https://canaireadme.netlify.app/api/v1/agent/profile        {"domain":"acme.com"}
POST https://canaireadme.netlify.app/api/v1/agent/fix-package    {"url":"acme.com"}   (402 with a quote until paid)
GET  https://canaireadme.netlify.app/business/acme.com.json      Business Passport v1, cached, never crawls
GET  https://canaireadme.netlify.app/api/v1/agent/pricing        prices and payment rails

curl -s https://canaireadme.netlify.app/api/v1/agent/audit -H 'content-type: application/json' -d '{"url":"acme.com"}'

Specs: OpenAPI 3.1, Business Passport v1 JSON Schema, llms.txt.

Business Passport v1

Every field says where it came from: observed on the public website, claimed by the business owner, or verified (today: proven control of the domain). It also says whether the website states it or it was inferred, how confident we are, the source URL and quote, what could not be determined, and how fresh the data is. Actions are marked discovered (a public link to follow) or native (CanAIReadMe performs it; none today).

Payments

Calling a tool or endpoint never charges anyone.
Paid results come back as a quote: price, what is delivered, refund policy, a purchase id and access token, and a Stripe checkout link.
The user reviews and pays on Stripe. The agent then calls again with the purchase id and token to receive the package.
Retries with the same idempotency key return the same purchase and link, so a flaky network can't buy twice.

Errors and limits

Errors carry a stable code (for example INVALID_URL, WEBSITE_BLOCKED, RATE_LIMITED, PAYMENT_REQUIRED), whether a retry can help, how long to wait and the next step. Free callers get 20 new analyses a day; cached results are unlimited within the rate limit.

Honesty guarantees

The score is computed by deterministic checks. A language model never produces the number.
Text quoted from websites is third-party data. We flag website text that tries to steer AI answers.
Domains whose owners opt out are not profiled.
CanAIReadMe never changes a website.

Need higher limits? API keys raise them.