ARC · AGENT CHECKOUT

Agent commerce adoption tracker

Live: n = 3,309 stores checked, as of 11 Oct 2026 (UTC).

Free, public counts from saved HTTP checks, updated hourly. The open dataset is a dated weekly snapshot, so its counts can differ. 3325 hosts observed; 22 latest attempts failed; 333 latest attempts blocked.

Public JSON → · Browse stores → · Protocol Matrix → · Research pages →

Checkout WebMCP

1,334 of 1,334 (100%)

Checkout exposes WebMCP tools (get/update/complete, buyer approval required). HTML detection only; no purchase completed.

Among 1558 Shopify stores checked: 0 no, 224 unknown, 0 not checked. Rate excludes unknown and not checked. Other platforms: not applicable.

By commerce platform

Coverage: platform measured for 2,714 of 3,309 checked stores (n per platform below). The other 595 were blocked at the homepage or last checked before platform detection; they get platform evidence on their next check.

Detected from public homepage signals; “Not detected” includes headless/custom stores. Numbers are ARC's directory sample, not the whole web. Unavailable and older checks are excluded. Each percentage shows its measured denominator; product data and robots use evidence from the same completed check.

PlatformStores measuredWebMCPStorefront MCPUCP MCPValid UCPCheckout linkCheckout WebMCPProduct data averageStrong product dataBlocks answer agents
Adobe Commerce / Magento580% (n=58)0% (n=58)0% (n=58)0% (n=58)0% (n=58)Not applicable45.4 (n=23)17.4% (n=23)0% (n=58)
BigCommerce430% (n=43)0% (n=43)0% (n=43)93% (n=43)0% (n=43)Not applicable46 (n=10)10% (n=10)2.3% (n=43)
Lightspeed10% (n=1)0% (n=1)0% (n=1)0% (n=1)0% (n=1)Not applicable70 (n=1)0% (n=1)0% (n=1)
Not detected8340% (n=834)2.2% (n=834)2.2% (n=834)2.3% (n=834)2% (n=834)Not applicable56.4 (n=284)35.2% (n=284)6.4% (n=832)
PrestaShop120% (n=12)0% (n=12)0% (n=12)0% (n=12)0% (n=12)Not applicable57.5 (n=6)16.7% (n=6)0% (n=12)
Salesforce Commerce Cloud1520% (n=152)0% (n=152)0% (n=152)0% (n=152)0% (n=152)Not applicable59.2 (n=87)32.2% (n=87)6.6% (n=152)
Shopify155893.1% (n=1558)94.2% (n=1558)95.7% (n=1558)95.8% (n=1558)81.2% (n=1558)100% (n=1334)70 (n=1467)37.5% (n=1467)0.7% (n=1557)
Squarespace40% (n=4)0% (n=4)0% (n=4)0% (n=4)0% (n=4)Not applicable61.7 (n=3)0% (n=3)0% (n=4)
Wix10% (n=1)0% (n=1)0% (n=1)100% (n=1)0% (n=1)Not applicable65 (n=1)0% (n=1)0% (n=1)
WooCommerce510% (n=51)0% (n=51)0% (n=51)0% (n=51)0% (n=51)Not applicable50 (n=33)12.1% (n=33)11.8% (n=51)

1,451 of 3,298 (44%)

Shopify WebMCP adapter detected

Of stores checked since WebMCP detection began (6 Oct 2026, 18:00 UTC)

1,519 of 3,309 (45.9%)

Storefront MCP answered (/api/mcp listed tools)

Of all stores with a completed check

1,543 of 3,309 (46.6%)

UCP MCP endpoint

Of all stores with a completed check

1,598 of 3,309 (48.3%)

Valid UCP profile

Of all stores with a completed check

1,311 of 3,309 (39.6%)

Agent checkout link returned

Of all stores with a completed check

334 of 3,309 (10.1%)

MCP blocked

Of all stores with a completed check

Google agentic checkout readiness

These are the public parts of Google's checklist that ARC can check from outside. This is not Google approval or certification. See Google's Merchant Center help on UCP / agentic checkout.

All public checks: 214 (11.2%) · 1912 fully measured stores

1,598 of 3,122 (51.2%)

Valid /.well-known/ucp profile

Of 3,122 stores where this item was measured (blocked or failed requests excluded), so this rate can differ from the headline count above

1,769 of 2,721 (65%)

Return policy findable

Of 2,721 stores where this item was measured (blocked or failed requests excluded), so this rate can differ from the headline count above

571 of 2,721 (21%)

Return window stated

Of 2,721 stores where this item was measured (blocked or failed requests excluded), so this rate can differ from the headline count above

1,666 of 2,721 (61.2%)

Shipping policy findable

Of 2,721 stores where this item was measured (blocked or failed requests excluded), so this rate can differ from the headline count above

1,497 of 2,721 (55%)

Customer support contact findable

Of 2,721 stores where this item was measured (blocked or failed requests excluded), so this rate can differ from the headline count above

1,513 of 1,928 (78.5%)

Product GTIN / MPN / SKU present

Of 1,928 stores where this item was measured (blocked or failed requests excluded), so this rate can differ from the headline count above

native_commerce(checkout_eligibility): Can't be checked from outside — set in your Merchant Center feed

  • Apparel: 10 (6.4%) pass all public checks / 156 measured
  • Beauty: 19 (9.1%) pass all public checks / 209 measured
  • Electronics & accessories: 9 (8.6%) pass all public checks / 105 measured
  • Food & drink: 23 (9.1%) pass all public checks / 253 measured
  • Footwear: 18 (11.8%) pass all public checks / 153 measured
  • Home: 18 (9.7%) pass all public checks / 186 measured
  • Jewelry: 23 (12.4%) pass all public checks / 185 measured
  • Kids: 23 (12.3%) pass all public checks / 187 measured
  • Other: 0 pass all public checks / 1 measured
  • Outdoor: 17 (11.8%) pass all public checks / 144 measured
  • Pets: 19 (13.8%) pass all public checks / 138 measured
  • Supplements & wellness: 35 (17.9%) pass all public checks / 195 measured

AI agent access in robots.txt

n = 2983 stores with measured robots evidence in their latest completed check. Unavailable and older rows are excluded. A block means the home or tested product path is disallowed; bots may ignore robots.txt, and firewalls may block allowed bots.

25 (0.8%)

No robots.txt (HTTP 404/410; refusals such as 403 are excluded)

102 (3.4%)

Blocking at least one answer agent

  • Bytespider (training): 87 blocked / 2983 measured (2.9%)
  • CCBot (training): 84 blocked / 2983 measured (2.8%)
  • GPTBot (training): 78 blocked / 2983 measured (2.6%)
  • Amazonbot (answer): 76 blocked / 2983 measured (2.5%)
  • Amazonbot-Train (training): 76 blocked / 2983 measured (2.5%)
  • ClaudeBot (training): 72 blocked / 2983 measured (2.4%)
  • Meta-ExternalAgent (training): 69 blocked / 2983 measured (2.3%)
  • Gemini-User (answer): 67 blocked / 2983 measured (2.2%)
  • Google-Extended (training): 67 blocked / 2983 measured (2.2%)
  • Bingbot (search): 57 blocked / 2983 measured (1.9%)
  • Copilot (answer): 57 blocked / 2983 measured (1.9%)
  • PerplexityBot (answer): 53 blocked / 2983 measured (1.8%)
  • ChatGPT-User (answer): 51 blocked / 2983 measured (1.7%)
  • Applebot (search): 49 blocked / 2983 measured (1.6%)
  • Claude-User (answer): 47 blocked / 2983 measured (1.6%)
  • Claude-SearchBot (search): 46 blocked / 2983 measured (1.5%)
  • DuckAssistBot (search): 46 blocked / 2983 measured (1.5%)
  • OAI-SearchBot (search): 46 blocked / 2983 measured (1.5%)
  • Perplexity-User (answer): 46 blocked / 2983 measured (1.5%)
  • Googlebot (search): 44 blocked / 2983 measured (1.5%)
  • Meta-ExternalFetcher (answer): 44 blocked / 2983 measured (1.5%)

Return & shipping policies agents can find

n = 2721 stores with measured policies in their latest attempt. From the store's pages and structured data; presence does not verify policy terms. Unavailable and legacy rows are excluded.

1769 (65%)

Returns findable

298 (11%)

Returns in structured data

571 (21%)

Stated return window

1666 (61.2%)

Shipping findable

240 (8.8%)

Shipping in structured data

Product data from the page's structured data

n = 1928 stores with measured product data in their latest check. Unavailable and older unmeasured rows are excluded. Server HTML JSON-LD only.

1513 (78.5%)

GTIN / MPN / SKU

1616 (83.8%)

Availability

195 (10.1%)

Return policy

184 (9.5%)

Shipping details

By category

Percentages appear only with at least 10 completed hosts. All rates use completed checks.

CategoryCheckedShopify WebMCP adapter detectedStorefront MCP answered (/api/mcp listed tools)UCP MCP endpointValid UCP profileAgent checkout link returnedMCP blockedLatest failedLatest blocked
Apparel30175 (24.9%)82 (27.2%)84 (27.9%)90 (29.9%)73 (24.3%)43 (14.3%)342
Beauty305158 (51.8%)168 (55.1%)171 (56.1%)176 (57.7%)153 (50.2%)32 (10.5%)132
Electronics & accessories28954 (18.7%)57 (19.7%)58 (20.1%)66 (22.8%)47 (16.3%)51 (17.6%)351
Food & drink399201 (50.9%)212 (53.1%)213 (53.4%)218 (54.6%)187 (46.9%)32 (8%)132
Footwear284122 (43.3%)129 (45.4%)131 (46.1%)137 (48.2%)99 (34.9%)25 (8.8%)125
Home350140 (40.5%)158 (45.1%)159 (45.4%)164 (46.9%)142 (40.6%)32 (9.1%)132
Jewelry281155 (55.2%)156 (55.5%)160 (56.9%)161 (57.3%)144 (51.2%)25 (8.9%)025
Kids279161 (57.7%)156 (55.9%)161 (57.7%)166 (59.5%)140 (50.2%)16 (5.7%)216
Other312222000
Outdoor279113 (40.5%)118 (42.3%)120 (43%)125 (44.8%)99 (35.5%)31 (11.1%)331
Pets260102 (39.2%)103 (39.6%)104 (40%)111 (42.7%)84 (32.3%)27 (10.4%)527
Supplements & wellness279169 (60.8%)178 (63.8%)180 (64.5%)182 (65.2%)141 (50.5%)20 (7.2%)220

Daily history

Latest completed checks as of each UTC day. Today is partial. This is a changing curated sample.

WebMCP adapter detection rate by day

2026-10-06 — 2026-10-11; scale 0–100%

Method note: WebMCP detection began at 18:00 UTC on 6 Oct 2026. Earlier checks are not measured for WebMCP and are excluded from its rate, so 6 Oct uses only checks after that time. An earlier version of this page showed 39.8% on 6 Oct rising to 54% on 7 Oct; that was the detector starting, not stores adopting WebMCP. Method changelog →

UTC dayCheckedShopify WebMCP adapter detectedStorefront MCP answered (/api/mcp listed tools)UCP MCP endpointValid UCP profileAgent checkout link returnedMCP blockedLatest failedLatest blocked
2026-10-06378149 (51.9%)240 (63.5%)242 (64%)247 (65.3%)221 (58.5%)32 (8.5%)432
2026-10-071399674 (48.7%)720 (51.5%)729 (52.1%)751 (53.7%)648 (46.3%)146 (10.4%)6146
2026-10-0833061448 (44%)1518 (45.9%)1542 (46.6%)1597 (48.3%)1310 (39.6%)334 (10.1%)16334
2026-10-0933081451 (44%)1518 (45.9%)1542 (46.6%)1597 (48.3%)1309 (39.6%)334 (10.1%)21333
2026-10-1033091450 (44%)1519 (45.9%)1543 (46.6%)1598 (48.3%)1311 (39.6%)334 (10.1%)22333
2026-10-1133091451 (44%)1519 (45.9%)1543 (46.6%)1598 (48.3%)1311 (39.6%)334 (10.1%)22333

Recent changes

  • optimumnutrition.com

    2026-10-10T04:41:41.000Z → 2026-10-11T05:11:44.000Z

    Shopify WebMCP adapter detected: no → yes

  • kickeepants.com

    2026-10-09T19:11:47.000Z → 2026-10-10T19:41:45.000Z

    Agent checkout link returned: no → yes

  • pixibeauty.co.uk

    2026-10-08T15:41:52.000Z → 2026-10-09T20:11:58.000Z

    Agent checkout link returned: no → yes

  • appaman.com

    2026-10-07T22:41:57.000Z → 2026-10-09T16:11:50.000Z

    Agent checkout link returned: yes → no

  • amazfit.com

    2026-10-07T22:41:41.000Z → 2026-10-09T15:41:59.000Z

    Agent checkout link returned: no → yes

  • sci-mx.co.uk

    2026-10-07T21:41:48.000Z → 2026-10-09T15:11:48.000Z

    Agent checkout link returned: yes → no

  • greatoutdoors.ie

    2026-10-07T17:41:44.000Z → 2026-10-09T11:11:43.000Z

    Agent checkout link returned: yes → no

  • florislondon.com

    2026-10-07T17:11:50.000Z → 2026-10-09T10:41:46.000Z

    Agent checkout link returned: no → yes

  • berghaus.com

    2026-10-07T15:11:45.000Z → 2026-10-09T08:12:04.000Z

    Agent checkout link returned: yes → no

  • 53degreesnorth.ie

    2026-10-07T14:11:42.000Z → 2026-10-09T07:12:01.000Z

    Shopify WebMCP adapter detected: no → yes

  • misen.com

    2026-10-06T23:41:45.000Z → 2026-10-09T03:41:43.000Z

    Shopify WebMCP adapter detected: yes → no

Methodology

Checkout WebMCP is a GET-only Shopify checkout HTML check introduced 8 Oct 2026. Rates use Shopify stores with measured checkout evidence; unknown failures and not-checked stores are separate from no. Tool names in initialization HTML are declarations, not a live agent session or completed purchase; completion requires buyer approval.

Platform rates use latest completed checks with measured homepage platform evidence; unavailable and legacy rows are excluded. Public homepage fingerprints cannot identify every headless or custom store (Not detected). Each metric has its own measured denominator. This is ARC’s directory sample, not the whole web.

Robots rates use measured robots evidence in the latest completed check per host, with their own denominator. 404, 410 and other 4xx mean no robots.txt. Refusals (401, 403, 407, 429, usually a bot firewall answering ARC's request), 5xx, network errors and rows saved before the HTTP status was recorded are excluded. Redirects are followed (up to 3). Blocks apply to the home or tested product path, not observed bot fetches. Bots may ignore robots and firewalls may block allowed bots. Named agents share some tokens across groups.

Google public readiness uses the latest attempt per host and separate measured denominators per item. All-public-checks requires every public item to be measured and pass. Legacy support evidence, blocked or failed UCP requests, and unavailable evidence are not measured. Merchant Center feed eligibility is excluded; this is not Google approval.

Policy rates use only hosts whose latest attempt has measured policies, with their own denominator. Findings are from the store’s pages and structured data; presence does not verify policy terms. Unavailable and legacy rows are excluded.

Product data rates use only hosts whose latest attempt has a measured productData result, with their own denominator. Unavailable and legacy rows are excluded. Fields are from the page’s structured data (server HTML JSON-LD only); presence does not verify accuracy.

The sample is ARC's curated directory and Index store list, plus hosts with saved public checks. It is not all of e-commerce. Each host is counted once.

Rates use the latest completed HTTP check per host. Failed checks are excluded from the denominator; an earlier completed check remains in use after a failed attempt. Failed hosts and blocked hosts are also reported from each host's latest attempt.

WebMCP detection began at 18:00 UTC on 6 Oct 2026. Checks saved before then are not measured for WebMCP and are left out of its denominator; the jump from 39.8% (6 Oct) to 54% (7 Oct) under the old counting was that detector start, not rollout. See /methodology/changelog.

WebMCP means Shopify's storefront adapter was detected in HTML, not a live agent session. Storefront MCP means the store's /api/mcp server answered a tools/list request with at least one tool (on Shopify today, the policy and FAQ tool; Shopify moved catalog and cart to UCP). UCP MCP means the UCP MCP endpoint was detected. A valid UCP profile passed validation at /.well-known/ucp. A checkout link means the cart tool returned a link for the tested product, not a completed purchase. MCP blocked means an observed block response.

Daily values use checks saved on or before the end of each UTC day; today is partial. Rates can change as the sample grows. Categories use the curated directory mapping, then the Index mapping; unknown hosts are Other. Category percentages are hidden below 10 completed hosts.

History comes from append-only saved checks. Changes compare each host's latest two completed checks. Reads make no store HTTP requests or model calls and do not change ARC scores.

Cite this: ARC Agent commerce adoption tracker, arcreport.ai/agent-checkout/adoption, accessed 11 Oct 2026.