ARC Agentic Report Card MCP¶

Hosted MCP service (Streamable HTTP, keyless for reads) - ARC, the Agentic Report Card, is a free and neutral record of which online stores AI agents can actually buy from, with dated evidence, plus readiness testing and fix lists for the stores themselves.

Server type: Remote (Streamable HTTP), hosted service at arcreport.ai
Auth: None for the record reads and checkout reporting (vendor stated); the monitor tools manage your store checks
Connectors: https://www.arcreport.ai/api/mcp (store checks, 19 tools) and https://www.arcreport.ai/api/ask (public index, 15 tools)
Registry: ai.arcreport/mcp and ai.arcreport/ask (official MCP Registry)
Pricing: Free; the underlying dataset is open (CC-BY-4.0) at arcreport.ai/data
Built by: ARC (arcreport.ai)
Listing: mcp.so feed, October 13, 2026 evening sweep

Why This Matters for Operators¶

Agent commerce turns "can a customer check out on my store" into "can an agent check out on my store", and that question is about to sit in front of every product conversation an assistant has. ARC keeps the checkable record: which stores support machine checkout, which protocols they speak, when the last verification ran, and which checks are failing.

For an e-commerce team the useful half is the second one. The readiness checks and the fix list turn "agents may not be able to buy from us" into a work queue: product data, checkout surface, protocol support. The monitors then keep watching, so the answer stays current instead of being a one-time audit.

Tools & Capabilities¶

Store checks connector (/api/mcp, 19 tools):

Tool	Purpose
check_store	Check a store for agent checkout reachability, with dated evidence
find_stores	Search the record for stores by reachability and result
can_agent_checkout	One-call answer for whether an agent can complete a checkout on a store
google_checkout_readiness	A store against Google's agent checkout requirements
product_data_check	Whether product data is ready for agent purchase
check_agent_readiness / get_readiness_result / list_readiness_checks	Run readiness checks and read the results and the check catalogue
get_fixes	The fix list for failing checks
get_store_hub	A store's hub: checks, history and evidence
ask_agents_about_brand	Agent-perspective lookups for a brand
report_checkout	Report a checkout attempt as dated community evidence
bulk_create_monitors / list_monitors / monitor_summary / update_monitor / pause_monitor / resume_monitor / run_monitor_now	The monitoring suite for store checks

Public index connector (/api/ask, 15 tools):

Tool	Purpose
ask_arc	Ask questions of the ARC index
find_agent_ready_stores / store_card	Find agent-ready stores and read shareable store cards
agent_commerce_adoption / agent_leaderboard	Adoption statistics and the readiness leaderboard
protocol_matrix	Which agent commerce protocols each store supports
validate_ucp_profile	Validate a UCP profile
get_agent_facts / list_category_agent_facts	Structured facts for a brand or a category
dataset_info	The open dataset: scope, licence, refresh cadence
submit_store / submit_leaderboard_run	Submit a store or a leaderboard run

Installation¶

claude mcp add --transport http arc https://www.arcreport.ai/api/mcp
claude mcp add --transport http arc-ask https://www.arcreport.ai/api/ask

Any MCP client that speaks Streamable HTTP can add the same URLs. Reading the record needs no sign-in.

Configuration¶

{
  "mcpServers": {
    "arc": {
      "type": "http",
      "url": "https://www.arcreport.ai/api/mcp"
    }
  }
}

Business Relevance¶

E-commerce and DTC teams test whether agents can complete a checkout on their store, get the fix list for what is broken, and monitor the result over time
Agencies and platforms check client stores in one call and compare protocol coverage before recommending changes
Agent builders find agent-ready stores to route purchases to, and read the protocol matrix and the adoption numbers
Brands and marketplaces submit their store, add dated evidence, and watch the leaderboard

Integration with CorpusIQ¶

ARC covers the agent-commerce layer: whether agents can buy, what is broken and what changed. CorpusIQ answers the surrounding business question from connected sources such as store data, spreadsheets, email and accounting, so a readiness conversation that starts in a product meeting can pull the checkout evidence into the same thread.

Used together: ask CorpusIQ for the operating view (orders, inventory, marketing, correspondence) and ask ARC for the agent-commerce view (reachability, protocol support, fixes). ARC stays the source of record for the checkout evidence; CorpusIQ remains the place the rest of the question is answered alongside.

Limitations¶

A store absent from the record is not proven non-buyable; coverage depends on the checks that have been run, so read the dates on the evidence
The monitor tools manage your store checks and may involve an account flow; the record reads and checkout reporting need no sign-in
The agent checkout space is young and moving fast; protocol support changes, so re-check before making claims to customers
Operated as a free service by ARC; the dataset is published openly (CC-BY-4.0)

FAQ¶

Is it really keyless?¶

Yes for the record: reading saved checks (check_store, find_stores) and reporting a checkout attempt need no sign-in. A live probe returned 200 with all 19 tools on the store checks connector and all 15 on the public index connector.

What can an assistant actually do with it?¶

Ask whether a store is agent-buyable and see the dated evidence, run readiness checks and read the fix list, compare protocol support across stores, read adoption and leaderboard numbers, and keep monitors watching a store over time.

Who is it for?¶

E-commerce and DTC operators, agencies that manage stores, agent builders that need to route purchases, and brands that want their store on the record.

What does it cost?¶

The record and the checks are free. The underlying dataset is open (CC-BY-4.0) at arcreport.ai/data.