ARC (arcreport.ai): can an agent reach checkout?
Server Details
Free open check: can an AI agent reach checkout on this store? Advisory; ARC stops before payment.
- Status
- Healthy
- Uptime
- 100.0% over the last 24 hours
- Last Tested
- Transport
- Streamable HTTP · MCP 2025-11-25
- URL
- ChatGPT
- –
- Claude
- –
TDQS
Scored across 7 tools
Each tool has a distinct primary call moment: pre-attempt check, multi-store comparison, discovery, post-attempt reporting, failure explanation, published record, and store-owner guidance. The descriptions explicitly route between before_checkout, why_blocked_and_fix, and store_record, though the overlap in returned store-verdict data creates minor ambiguity.
All names use snake_case and are readable, but conventions are mixed: some are verb_noun (find_stores, report_outcome), others are noun/prepositional phrases (before_checkout, store_record) or a full phrase (why_blocked_and_fix). The set is understandable but not a predictable pattern.
Seven tools is well within the typical 3-15 range for a focused advisory server. Each tool maps to a clear workflow stage and earns its place without obvious redundancy.
The surface covers pre-check, comparison, discovery, post-attempt logging, failure diagnosis and fix lookup, published record retrieval, and store-owner guidance. For an advisory checkout-readiness domain, there is no obvious lifecycle gap.
Available Tools
7 toolsbefore_checkoutBefore checkoutARead-onlyIdempotentInspect
Call this right before your agent tries to check out at a store, or when a user asks "can my agent check out at ?". Not for an attempt that already happened (log it with report_outcome), the reason an agent keeps failing (why_blocked_and_fix), or choosing between 2 or more stores (compare). Returns a one-line answer, reaches_checkout (true only when a checkout link came back or a shopper reached checkout; false when a route was measured and did not reach checkout; null when not shown either way: no agent catalog found and no browser test is null, and so is a product added to the cart with no shopper reaching checkout; null is not a no), protocol_route, browser_route, standard_shoppers when ARC's standard shoppers tried the store (a dated label, "ARC's standard shoppers: N of M reached checkout, ": a lower bound, not a verdict; ARC stops before payment and never places an order), best_route (the best saved way in), payment_methods_declared (the payment handler ids the store declares to agents in its own UCP profile, with source "ucp_profile", checked_at and a note; the store's own declaration, not a test of any agent paying there; null with payment_methods_note when ARC has no UCP profile reading, which is not a no), up to 3 blockers, checked_at, the Agent Readiness Grade (grade, with expired when over 30 days old; grade_note when not graded, which is not a mark against the store), alternatives_hint when the grade is D or F or expired (a find_stores call with min_grade B in the same category), and a full_record link (the precheck answer, which also carries product_markup). Example: before_checkout({"url":"https://allbirds.com"}) → {"answer":"A saved check got a checkout link back at allbirds.com via storefront_mcp (6 hours ago).","reaches_checkout":true,…}. The answer leads with the dated fact; once the record is not current it starts "Last known on .". reaches_checkout false or unknown is not a no. Saved checks only: no store fetch, no scan, no login. Advisory; ARC stops before payment. Every answer carries checked_at (the saved check it rests on; for a list, the oldest) and stale (true when that check is 24 hours old or older, or has no time). Errors: {error:{code, message, next_step}}.
Inputs: url (required).
compareCompare storesARead-onlyIdempotentInspect
Call this when a user names 2 or more stores for the same item and asks which one to use, or when you must choose between a few stores and want the one where an agent is most likely to reach checkout. Use it instead of calling before_checkout once per store. Pass 2 to 10 store domains or URLs. Returns one short row per store (reaches_checkout, best_route, checked_at, grade, payment_methods_declared as on before_checkout), and pick: among stores whose saved check reached checkout in the last 30 days, the one with the best current Agent Readiness Grade (expired after current of the same letter, not graded last); null when no store has such evidence. A row with stale: true is more than 24 hours old: re-check right before the attempt. Rows are ordered only by saved checkout evidence, its freshness and published grade, never by anything else. Example: compare({"stores":["allbirds.com","gymshark.com"]}) → {"pick":"allbirds.com","stores":[…]}. Each store counts as one precheck read. Saved checks only: no store fetch, no scan, no login. Advisory; ARC stops before payment. Every answer carries checked_at (the saved check it rests on; for a list, the oldest) and stale (true when that check is 24 hours old or older, or has no time). Errors: {error:{code, message, next_step}}.
Inputs: stores (required).
find_storesFind storesARead-onlyIdempotentInspect
Call this when you need to pick a store for a product need and want one where a saved check shows an agent reached checkout. Filters: category or q (shopper words), min_grade (optional, default none), payment_method (optional: a payment handler id such as com.google.pay, or "google pay" or "shop pay"; keeps stores whose saved UCP profile declares it), limit (default 5). Returns total, the store hosts with checked_at and grade, and a full_list link; with payment_method, also payment_method (id, name, declared_by: how many listed stores declare it, and a note: a declaration is the store's own statement, not a test of any agent paying there). Each store carries its Agent Readiness Grade (grade, null when not graded). Stores whose saved checkout evidence is from the last 30 days come first (fresh_success: true), then older evidence. Within each, best current grade first (A, then B…); an expired grade ranks after current ones of the same letter; newest checked_at first within a grade. Results are ordered only by published grade and freshness, never by anything else. Pass min_grade to keep only stores with a current grade of that letter or better. Example: find_stores({"category":"running shoes","limit":3}) → {"total":12,"stores":[{"host":"…","checked_at":"…","grade":{"grade":"A",…}}],…}. Example: "Which stores declare Google Pay to agents?" → find_stores({"payment_method":"google pay"}) → {"payment_method":{"id":"com.google.pay","declared_by":…},"stores":[…],…}. An unknown payment_method is error code unknown_payment_method with handler_ids (the ids ARC has seen); not_available means retry in a minute. A store missing from the list is not a no. Saved checks only: no store fetch, no scan, no login. Advisory; ARC stops before payment. Every answer carries checked_at (the saved check it rests on; for a list, the oldest) and stale (true when that check is 24 hours old or older, or has no time). Errors: {error:{code, message, next_step}}.
Inputs: q, limit, category, min_grade, +1 more.
guideMerchant guideARead-onlyIdempotentInspect
Call this when a person at a store (a CEO, a marketing or e-commerce manager, or a developer) asks where their store is on the journey to AI assistants, what to do next on their platform, or how stores in their category compare. Not for an agent about to check out: use before_checkout for that. Returns the store's place on ARC's journey map (arc-journey/v1: 5 stages, each met, not met or "Not measured" with its date; not measured is not a no, and there is no score or grade), the next 3 steps for its platform at the chosen depth with source links and an as-of date (a platform fact ARC has not sourced says "Not yet documented by ARC"), how many stores ARC checked in its category and on its platform and how many meet each stage (counts with denominators and the snapshot date), and up to 5 named stores that meet the furthest stage (ordered only by stage reached, then the newest reading). role: ceo, manager or developer sets depth (basics, detail or implement) unless depth is given. Send platform alone for the platform guide with no store. Example: guide({"store":"gymshark.com","role":"manager"}) → {"answer":"gymshark.com: Meets stages 1, 2, 3 and 4. Not met: stage 5. Saved check 16 Oct 2026.","stages":[…],"next_steps":[…],"comparison":{…},"examples":{…}}. Saved checks only: no store fetch, no scan, no login. Advisory; ARC stops before payment. Every answer carries checked_at (the saved check it rests on; for a list, the oldest) and stale (true when that check is 24 hours old or older, or has no time). Errors: {error:{code, message, next_step}}.
Inputs: role, depth, store, category, +1 more.
report_outcomeReport the outcomeAInspect
Call this right after your agent tried to check out at a store, to log, record or report what happened: it reached checkout, a captcha stopped it, it got as far as shipping, anything. Use it for any attempt that already happened, so the next agent knows; before_checkout is only for before an attempt. Free, no login. Omit contributor_key and ARC issues one; send it on later reports. No personal data: emails, phones, addresses, card numbers and IPs are rejected or stripped. Example: report_outcome({"store_domain":"allbirds.com","agent_family":"browser","agent_name":"my-agent","timestamp":"2026-11-05T10:00:00Z","furthest_step":"checkout","blocker":"none"}) → {"ok":true,"status":"unverified",…}. A report stays private until 3 independent contributors agree. Same as POST /api/outcomes and report_checkout on /api/mcp. checked_at and stale are null on this answer: a report is not one of ARC's checks. Errors: {error:{code, message, next_step}}.
Inputs: blocker (required), agent_family (required), store_domain (required), furthest_step (required), +4 more.
store_recordStore recordARead-onlyIdempotentInspect
Call this when a user asks what ARC recorded for a store, or you need the citable store page. Returns the store page's verdict, checked_at (the day of the check), stale, category, tried (ARC's standard shoppers' newest dated run, or "Not measured"), the recorded fix, payment_methods_declared (as on before_checkout; null with payment_methods_note when ARC has no UCP profile reading, which is not a no), record_url (claimed, served, date, method and product_markup as JSON) and links to the page, its JSON and Markdown. Example: store_record({"store":"allbirds.com"}) → {"verdict":"…","page_url":"https://www.arcreport.ai/stores/allbirds.com",…}. Reads the published snapshot behind /stores/. Saved checks only: no store fetch, no scan, no login. Advisory; ARC stops before payment. Every answer carries checked_at (the saved check it rests on; for a list, the oldest) and stale (true when that check is 24 hours old or older, or has no time). Errors: {error:{code, message, next_step}}.
Inputs: store (required).
why_blocked_and_fixWhy blocked, and the fixARead-onlyIdempotentInspect
Call this after before_checkout says reaches_checkout is false or null, or when an agent got stuck at a store and you want the reason and the fix. Returns why (the saved blockers), fix (the store page's recorded fix, or "not recorded"), next_step and links to the full record. Example: why_blocked_and_fix({"store":"gymshark.com"}) → {"why":["…"],"fix":"…","next_step":"…"}. Saved checks only: no store fetch, no scan, no login. Advisory; ARC stops before payment. Every answer carries checked_at (the saved check it rests on; for a list, the oldest) and stale (true when that check is 24 hours old or older, or has no time). Errors: {error:{code, message, next_step}}.
Inputs: store (required).
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
Related MCP Connectors
Can an AI agent buy on this French store? 500 stores tested: score, guest checkout, UCP. Free.
Free, no sign-in MCP: can AI agents reach checkout at a store? Try: Use ARC to check allbirds.com
AgentReady.market audit: can an AI shopping agent find, understand and BUY on this store? /100.
Is a website ready for AI shopping agents? Readiness score (0-100) + agent shopping simulation.
Related MCP Servers
- AlicenseAqualityBmaintenanceFree, rules-based checks for e-commerce product pages: AI shopping readiness (structured data, reviews, return and shipping terms, GTINs, AI crawler access), Google Merchant Center compliance, store SEO audit and Shopify checkout deadlines. No AI calls, nothing stored.24MIT
- AlicenseNot gradedqualityCmaintenanceEnables AI agents to check whether online stores are reachable at checkout using dated, source-labelled evidence, find stores, and report checkout attempts through hosted Streamable HTTP tools.MIT
- FlicenseNot gradedqualityDmaintenanceEnables AI agents to browse product catalogs, search products with filters, and initiate checkouts, generating order summaries and checkout URLs.-

CartAI MCP Serverofficial
AlicenseNot gradedqualityCmaintenanceEnables AI hosts to search products, retrieve variants and pricing, estimate shipping and tax, authorize payments, store merchant credentials, and autonomously execute and track checkouts across merchants, including interactive MFA support.1Apache 2.0
Glama MCP Gateway
Add one secure layer between your agents and this server.