Vercel's Is Agentic score — and what a store-only audit adds
Searching for your Is Agentic score? is-agentic.com, from Vercel, scores any public website "based on what agents can discover, access, and use", and "every scan is run by Ora". Its CLI (npx is-agentic), public JSON API and MCP server are free and need no API key, per Vercel's own skill file. AgentReady, at agentready.market, is an independent audit with a narrower subject: the path an AI shopping agent takes from an online store's front door to its checkout, scored by our own versioned engine, the same way on every platform, re-scannable on demand (free scans come with a monthly allowance; Pro is unlimited and re-scanned daily) and ranked in a public index of stores.
The verdict
Is Agentic is a broad agent-readiness score for any site (docs, apps, businesses, stores), and it shows one observed agent journey next to the number. AgentReady scores one kind of site, the store, on the steps of an order, with every check a rule over what the store returned. Use Is Agentic to see how agents use your site in general; use AgentReady to see whether a shopping agent can get into your store, read the offer and find the buy action, and how you rank against other stores.
Last reviewed 2026-10-09. Is Agentic's column quotes is-agentic.com, its methodology page, Vercel's is-agentic agent skill and Ora's methodology page, all read on the date above; where they don't say something, neither do we. We did not run their scanner for this page.
Best for
- Is Agentic — any type of site (it has documentation and content, business, app and commerce views), developers who want the score in a CLI, an API or an MCP host, and anyone who wants to see one observed agent journey through the site
- AgentReady — online stores that need the order path audited — shopping-agent access, offer data, variants, returns and shipping, the rails, the buy action — with a score from a versioned engine and a rank among other stores
Side by side
| What gets checked | Is Agentic | AgentReady |
|---|---|---|
| Scope | Any public website, domain or public MCP endpoint, per Vercel's skill; new root-site scans have documentation and content, business, app and commerce views | Built for e-commerce stores, on any platform |
| Who scores it | Ora: "Every check is run and scored by Ora"; Is Agentic groups the results | Our own engine; the scoring version is stored with every scan |
| How the score adds up | Essential checks 80 points, Recommended 20, a bonus capped at 5. Ora's own scale, on ora.ai: Discovery 20, Access 30, Usability 40, Payments 10 | Four weighted pillars: access 34, structured data 30, catalog 20, transaction rails 16; a robots.txt that blocks every AI crawler we check, or a firewall or CDN that turns one away, caps the score at 25 |
| Where the checks come from | Ora spawns "real agents across thousands of intents and sites"; each check is "reverse-engineered from what actually moved an agent's success" | Written for each step of an order; weights and rails published on /methodology, every check named in each report |
| Commerce and payments | Recommended checks activate when the scan identifies a commerce surface; Ora's Payments layer asks "can it pay you?". The pages we read don't list the product or offer fields checked | Scored check by check: Product JSON-LD, price and currency, availability, brand, identifiers, variants and attributes; return policy and shipping in the offer; the UCP, A2A and ARD rails |
| Bot blocks | Ora's Access layer names "blocked crawlers" among what goes wrong. Is Agentic's homepage, on what a report describes, names status codes and crawler policies among the evidence that can change after the snapshot, and notes that bot defenses can produce behavior a public scan cannot observe. The pages we read don't say whether it fetches as named AI user agents | Your homepage fetched as AI crawlers (compared with a browser fetch) and as shopping agents (ChatGPT-User, Claude-User, Perplexity-User, Meta-ExternalFetcher, AmazonBuyForMe; judged on the response status) |
| An observed agent journey | Yes: each report includes an observed agent journey showing how one agent navigated the site, shown as supporting evidence and not part of the score | No: no agent and no model in the scan, so there is no journey to watch; the same responses from the store give the same score |
| What one report covers | Reports are keyed to one exact public URL (example.com and example.com/docs are separate reports, per Vercel's skill). Its homepage says a report describes evidence available from one public URL at a particular time, and names status codes, crawler policies, rendered content, documentation, advertised APIs and emerging agent protocols as evidence that can change after the snapshot. The pages we read don't list which resources a scan fetches | The URL you give, robots.txt, a product page found from it, the agent files and the cart and checkout doors. A separate Pro catalog-coverage run samples up to 8 product pages across the catalog |
| Freshness | Rescan on demand from the report page; otherwise results may come from Ora's six-hour cache, stale reports may refresh in the background, and snapshots may be kept for historical comparison. A score can change when the site, Ora's methodology or a time-sensitive check changes | A scan you start runs fresh; Pro stores are re-scanned daily; the Index shows each store's latest scan |
| Public scores / ranking | Featured scores, and recent scores that rotate hourly, on its homepage. Ora, which runs the scans, links a Leaderboard from its site; the pages we read don't say whether Is Agentic scans appear on it | Scanned stores ranked in the AgentReady.market Index, plus a dated monthly census with open data (/census) |
| Developer access | Free CLI (npx is-agentic), public JSON API and MCP server, no API key | API and MCP server, documented at /api |
| Price | CLI, API and MCP server free, per Vercel's skill; the pages we read state no price for the web scan | Scan free · Pro on /pricing |
FAQ
What does the Is Agentic score measure?
In its own words, "the public parts of a website that AI agents can discover, retrieve, understand, and use". Essential checks share 80 points and cover fundamentals such as server-rendered content, correct HTTP behavior, clear document structure, recoverable errors, usable controls and identity metadata (a canonical URL, a declared language, Open Graph tags). Recommended checks share 20 points and switch on when the scan finds an API, OAuth flow, GraphQL endpoint, MCP server, developer portal or commerce surface. Emerging formats add a bonus capped at five points, and their absence never lowers a score.
Does Is Agentic check commerce and payments?
Yes, in its own way. Recommended checks activate when the scan identifies a commerce surface, its methodology has a commerce site-type view, and Ora's methodology scores a Payments layer ("can it pay you?", 10 of Ora's 100 points). The pages we read don't say which product or offer fields those checks read, so we can't compare them field by field. AgentReady scores those fields one by one: price, availability, identifiers, variants, returns and shipping.
Why can an Is Agentic score and an AgentReady score disagree?
They measure different things, scored differently. Is Agentic's score comes from Ora's checks, and its methodology says a score can change when Ora's methodology changes; its agent journey is kept out of the score because "one task cannot represent every agent, user goal, or production condition". AgentReady runs no agent: each check is a rule over what the store returned (status, headers, HTML, timing), versioned in our engine. More on why scanners diverge: /research/why-agent-readiness-scores-disagree.
Which one should I use?
For a docs site, an app or an API, Is Agentic: it has documentation and content, business and app views, and checks that switch on when it finds an API, OAuth flow, GraphQL endpoint or MCP server; our checks are built for stores only. For an online store, the two answer different questions: Is Agentic for how agents navigate and use the site, AgentReady for whether a shopping agent can get in and find where an order starts, and how the store ranks among others. Our scan stops at the checkout boundary and never places an order; the Is Agentic pages we read don't say which task the observed journey attempts.
Is AgentReady affiliated with Vercel, Ora or Is Agentic?
No. AgentReady at agentready.market is operated by Villepinte Studio (NL). It is unaffiliated with Vercel, Ora and is-agentic.com, and with the other products using the "AgentReady" or "Agent Ready" name (see /vs/namesakes).
Sources: Is Agentic · Is Agentic methodology · Vercel's is-agentic agent skill · Ora: how scoring works · Five commerce scanners compared (our research) · Our scoring methodology (deterministic, versioned)
More comparisons: vs Cloudflare's score · vs Shopify's audit · vs DIY SEO tooling · vs the namesakes · the best agent-readiness tools (2026) · all comparisons
AgentReady at agentready.market, operated by Villepinte Studio (NL) — unaffiliated with every other product named "AgentReady" or "Agent Ready", and with any third-party tool on this page. About us · the only authoritative pricing · contact.