Ask whether agents choose a brand — and whether they can reach checkout.
Agents are the customers.
Can an AI agent check out from any store?
Find stores where an AI agent can reach checkout: call find_agent_ready_stores on /api/ask with {"category":"running shoes","limit":10}. JSON: GET /api/agent-checkout/stores?category=footwear. Only real HTTP checkout-positive checks from the last thirty days are returned. One tested product; no purchase completed.
Free, no login, on /api/mcp: check_store with a domain, find_stores with a platform, protocol, or category, and product_data_check with a host or URL. They read saved checks. report_checkout sends one checkout attempt and stays private until 3 independent contributors agree. Spec: Outcome report. A live check is can_agent_checkout on /api/ask. GET /api/mcp is not a tool list. A new scan, fixes, and monitoring on /api/mcp need sign-in. google_checkout_readiness on this server needs sign-in; the same read on /api/ask does not.
GET https://www.arcreport.ai/api/agent-checkout?url=https://allbirds.com
{"name":"can_agent_checkout","arguments":{"url":"https://allbirds.com"}}Returns catalog, product, checkout-link, UCP, Shopify WebMCP adapter detection, and block facts, a fix, and a shareable result URL. Cached per host for 24 hours. Limits per network per UTC day: 50 requests and 10 fresh checks. Site-wide: 5,000 requests and 500 fresh checks. No order is placed. No score changes.
Try the free checkout check →▸ you: what do personal agents think of Allbirds, and can they reach checkout for the runners?
agent → calling ARC · ask_agents_about_brand
agent → calling ARC · check_agent_readiness
All without leaving the chat.
Connect it, then ask for a dated check.
Paste the free MCP URL below. No login. Reading the Index is also free at /for-agents.
Open your client settings
Open your MCP client’s Settings.
Add a connector
Add a remote MCP server in your client.
Paste the URL
Name it ARC, paste the URL below, click Add.
https://www.arcreport.ai/api/askThen ask: “Can personal agents reach checkout for this product, and did they choose us?”
Install configs for Claude Desktop, Claude Code, Cursor, and VS Code →
Quick start: precheck (free, no sign-in, CC BY 4.0)
Before your agent tries a store, ask ARC whether agents can reach checkout there.
curl -sS -A "arc-precheck/1.0" -H "X-ARC-Client: quickstart-curl" "https://www.arcreport.ai/api/precheck?url=https://absorbine.com&ref=quickstart-curl"
Example response (captured 2026-10-14, UTC; not live)
Run the curl for a current answer.
{
"host": "absorbine.com",
"last_checked": "2026-10-14T00:41:55.000Z",
"stale": false,
"known": true,
"reachable_checkout": true,
"rails": [
{
"name": "UCP",
"status": "valid",
"proves": "The saved check found a UCP profile that passed validation. That alone does not show an agent can reach checkout. Does not prove an agent completed checkout."
},
{
"name": "ACP",
"status": "not_found",
"proves": "The saved check did not find an ACP signal. Does not prove an agent completed checkout."
},
{
"name": "Shop Pay",
"status": "detected",
"proves": "The saved UCP file listed a Shop Pay payment handler. Does not prove an agent completed checkout."
},
{
"name": "WebMCP",
"status": "yes",
"proves": "The saved check detected Shopify's WebMCP adapter in storefront HTML. That is not a live browser session. Does not prove an agent completed checkout."
},
{
"name": "MCP",
"status": "yes",
"proves": "The saved check got an answer from a storefront or UCP MCP endpoint. Does not prove an agent completed checkout."
},
{
"name": "Checkout WebMCP",
"status": "detected",
"proves": "The saved Shopify checkout HTML declared Checkout WebMCP tools. Buyer approval is still required. No order is placed. Does not prove an agent completed checkout."
}
],
"blockers": [],
"confidence": "high",
"licence": "CC BY 4.0"
}reachable_checkout is based on ARC's last saved check. It does not mean an agent completed checkout. Check last_checked and stale before relying on it.
Contact: arcreport@mail.grokbot.com
Advanced server (19 tools)
Copy a config, then ask for a dated check.
Server URL: https://www.arcreport.ai/api/mcp. check_store, find_stores, and product_data_check are free and read-only. They return saved checks and the date ARC recorded them. They do not start a scan. A new scan, fixes, and monitoring need sign-in.
Use ARC to check allbirds.comClaude Desktop
Add custom connector: Settings → Connectors → Add custom connector. Paste https://www.arcreport.ai/api/mcp and leave authentication off. The free tools need no login.
The local claude_desktop_config.json file is a fallback. It accepts a command, not a URL:
{
"mcpServers": {
"arc": {
"command": "npx",
"args": ["-y", "mcp-remote@0.14.3", "https://www.arcreport.ai/api/mcp"]
}
}
}Claude Code
claude mcp add --transport http arc https://www.arcreport.ai/api/mcp
Cursor
Project file .cursor/mcp.json, or ~/.cursor/mcp.json.
{
"mcpServers": {
"arc": { "url": "https://www.arcreport.ai/api/mcp" }
}
}Add ARC in Cursor · or copy the install link
cursor://anysphere.cursor-deeplink/mcp/install?name=arc&config=eyJ1cmwiOiJodHRwczovL3d3dy5hcmNyZXBvcnQuYWkvYXBpL21jcCJ9VS Code
Workspace file .vscode/mcp.json, or the user mcp.json from MCP: Open User Configuration.
{
"servers": {
"arc": { "type": "http", "url": "https://www.arcreport.ai/api/mcp" }
}
}Claude is a trademark of Anthropic, Cursor of Anysphere, VS Code of Microsoft. ARC is not affiliated with or endorsed by them.
Free, no login
check_storeRead one store's saved check. Free, no login, read-only. Dates in the result are when ARC recorded the check. This does not start a scan.
find_storesSearch saved checks by platform, protocol, or category. Free, no login, read-only.
report_checkoutSend one checkout attempt. Free and no login. This is not a read. It stays private until 3 independent contributors agree.
product_data_checkRead the latest saved product data for a host or URL. Free, no login, read-only. A missing result is not a zero.
Sign-in
These need a signed-in ARC account on this server. They can start a scan, call a model, or change a monitor.
- get_store_hub
- google_checkout_readiness
- can_agent_checkout
- check_agent_readiness
- get_readiness_result
- list_readiness_checks
- ask_agents_about_brand
- get_fixes
- bulk_create_monitors
- list_monitors
- monitor_summary
- update_monitor
- pause_monitor
- resume_monitor
- run_monitor_now
Ask in plain English.
Benchmark personas for Muse, Instinct, Grok Bot, dots, and OpenClaw personas: did they mention you, recommend you, or name a competitor instead?
Can a personal agent reach the product page, understand it, and reach checkout? Returns a check id.
The score, the wall, the screenshots, and how many personal agents added the product to the cart. Same id does not start a new check.
Concrete next steps to get recommended, to be understood, and to reach checkout.
Past checks for this workspace, so you can reopen one without starting another.
See what personal agents would say — and whether they can reach checkout.
Connect ARC to the agent you already use, or run a scan on the web. Perception answers come from benchmark personas. Whether agents reach checkout is two lanes: a real agent-catalog check, and five browser cart shoppers. The catalog check does not change the score.