canagentsbuyfromyou

Cabfy for AI agents

Your agent-readiness audit, right inside your assistant.

Connect Cabfy to Cursor, Claude Code, Codex, or any compatible MCP client. Ask a question in plain English, run the right audit, and get useful results without leaving your workflow.

Connect with MCPView API documentation

Works with Codex, Claude Code, Cursor, and compatible remote MCP clients.

MCP

Start with a question

Your assistant selects the tools it needs and returns an actionable answer.

Status

“What tier is my site on?”

Priorities

“What is keeping me from the next tier?”

Comparison

“Compare my sites and find the most ready one.”

Connect

One URL, one key, and your assistant is ready.

Cabfy's MCP server uses the modern Streamable HTTP transport. Create an API key in your account, then add this URL as a remote server in your client.

  1. 01Create a Cabfy API key in your account.
  2. 02Add a remote MCP server with the URL above.
  3. 03Configure the Bearer token with your CABFY_API_KEY variable.
  4. 04Ask your assistant to analyze a domain authorized by your account.
Illustrative configuration →

MCP server URL

https://api.canagentsbuyfromyou.com/mcp/

Your key stays in your client. Never share it in a conversation or public repository.

Illustrative configuration

{
  "mcpServers": {
    "cabfy": {
      "url": "https://api.canagentsbuyfromyou.com/mcp/",
      "headers": {
        "Authorization": "Bearer ${CABFY_API_KEY}"
      }
    }
  }
}

Available tools

The full analysis cycle in one connector.

Expensive audits run asynchronously: launch them, then ask your assistant to track their status and read the result.

run_agent_readiness_scan

Launches the general scan without relying on a commerce protocol.

For discoverability, readability, and agent access.

run_agent_commerce_audit

Launches the commerce audit, with optional agent simulation.

For catalog, discovery, and agentic purchase-flow checks.

run_agent_performance_audit

Launches the PageSpeed-based Agent Performance audit.

For measuring an agent's browsing experience.

run_web_findability_audit

Launches the brand Web Findability audit.

For checking whether official resources can be found.

get_agent_audit_result

Retrieves an audit status and result by identifier.

Call after a launch to follow processing.

get_agent_performance_result

Retrieves the latest performance result for a domain.

For reading the most recent measurement.

get_web_findability_result

Retrieves the latest Web Findability result.

For reviewing sources found and priorities.

get_account_usage

Retrieves the connected account plan and quotas.

To check availability before launching an audit.

Why Cabfy + MCP

Go from question to decision without changing tools.

An independent engine

Your assistant uses the same checks and results as the Cabfy dashboard, not an opinion generated on the fly.

Contextual answers

Ask for a summary, comparison, or next action: the agent reads Cabfy data and explains it in your context.

Controlled access

The API key ties every call to your account, quotas, and audits. Revoke it from your account at any time.

You stay in control.

Cabfy launches and reads audits: no tool completes an order, triggers a payment, or changes your site. Calls use a Bearer API key and your plan's quotas.

Frequently asked questions

What is MCP?

MCP (Model Context Protocol) is a standard that lets an AI assistant discover and call external tools in a structured way.

Which assistants are compatible?

Any client that accepts a remote MCP server over Streamable HTTP, including Codex, Claude Code, and Cursor. Setup screens vary by client.

Can I run an audit without an account?

No. The MCP server requires a Cabfy API key so it can enforce account access, quotas, and result ownership.

Can my assistant buy or change my site?

No. Cabfy tools launch and read audits. They do not make payments or expose site-management actions.

Open the full documentation →Back to the landing →