UCP conformance score, top priorities, impact & effort matrix, and copy-paste fixes for every issue.
gravelmanseries.com publishes a conforming manifest with broad capability coverage.
Agent Discovery is solid (100/100).
Fix broken spec or schema URLs → Publish keys at the manifest root → Publish a schema URL for the service
One or more spec/schema URLs declared in the manifest do not resolve.
Agents have nothing to verify your responses against.
Service is missing its "schema" URL.
Prioritised recommendations by expected impact and implementation effort
High impact, low effort
High impact, high effort
Medium impact & effort
Lower priority
Response time benchmarked against 7,415 Wixes we monitor.
All improvements ranked by priority, with evidence and actionable fixes.
One or more spec/schema URLs declared in the manifest do not resolve.
Spec and schema URLs let agents validate request/response shapes before sending. When they 404 or time out, runtime validation degrades to "best effort" and strict agents reject the integration entirely. Common cause: copy-pasting a placeholder URL during initial manifest setup. See the Full Check Breakdown for the specific URLs that failed.
Agents have nothing to verify your responses against.
Signing keys let agents check that checkout responses really came from you. Tokenized payment flows in v2026-04-08+ require them. Since 2026-07-12 the spec expects a root-level "keys" list (the older "signing_keys" field is still accepted while stores move). Stores that hand the buyer a checkout link instead of taking payment through the agent can go without them for now.
Generate an Ed25519 keypair, publish the public key as a JWK in a root-level "keys" array (the canonical field since spec PR #566; "signing_keys" is legacy). Rotate annually; keep "kid" stable across rotations of the same key.
{
"keys": [
{
"kty": "OKP",
"crv": "Ed25519",
"x": "<base64url-encoded-public-key>",
"kid": "key-2026-04",
"alg": "EdDSA"
}
],
"ucp": { ... }
}Service is missing its "schema" URL.
The schema lets agents validate request/response shapes before sending. Highly recommended for any non-trivial endpoint.
"schema": "https://yourstore.com/.well-known/ucp/schemas/shopping.json"Every signal we evaluate, grouped by category.
Can AI agents find your store and its UCP manifest?
Manifest served over a secure connection.
Manifest URL uses https://
Agents reject non-HTTPS endpoints outright. Signed payloads over HTTP are meaningless because the channel itself is tamperable.
Manifest fetched OK from /.well-known/ucp.
HTTP 200
Agents start every session by fetching this URL. If it 404s or times out, the store is invisible to agent commerce.
robots.txt allows agent crawlers to fetch the manifest.
robots.txt allows /.well-known/ucp
Agents respect robots.txt. Even a published manifest gets ignored if the file is disallowed for known agent user-agents.
Hand-written brief at /llms.txt for AI agents.
Found at /llms.txt
Emerging convention (analogous to robots.txt). Recommendation/discovery agents read it for brand context, return policy, and crawl preferences.
XML sitemap published at /sitemap.xml.
Found at /sitemap.xml
Tells agents which pages are canonical and how often they change. Without it, agents fall back to following internal links.
OG meta tags on the homepage.
og:title and og:type/url present
OG tags drive how your store unfurls in agent surfaces, Slack, iMessage, and ChatGPT shares.
Schema.org Organization (or OnlineStore) JSON-LD on homepage.
Organization-typed JSON-LD detected
Gives agents a structured representation of your brand identity — used to verify you are who your manifest claims you are.
Viewport meta tag for mobile rendering.
<meta name="viewport"> present
Mobile agent sessions render in browser sandboxes that respect this. Without it, screenshots taken by agents are unreadable.
Does your manifest meet the published spec?
Manifest structure conforms to the published UCP spec.
No structural errors
Strict UCP agents reject manifests with structural errors. The error code in the diagnostic shows which field failed.
ucp.version follows YYYY-MM-DD format.
Version: 2026-08-25
UCP versions are dates, not semver. Agents pick the closest matching schema based on this string.
Root-level keys array (canonical since 2026-07-12; legacy signing_keys also accepted) for response verification.
No signing_keys at root (common for managed/redirect-checkout stores)
Required for tokenized payment flows in v2026-04-08+. Stores using checkout-link redirects (e.g. WooCommerce native) often skip this safely.
Manifest passes structural and field-level rules.
2 warnings
Each warning is a strict-spec deviation — the recommendations section above lists them with copy-code fixes.
Resolves the 12 spec/schema URLs declared across the manifest to confirm agents can fetch each one.
Did not resolve (6 of 12): https://ucp.dev/services/shopping/rest.openapi.json, https://ucp.dev/schemas/shopping/cart.json, https://ucp.dev/schemas/shopping/catalog-lookup.json, https://ucp.dev/schemas/shopping/checkout.json, https://ucp.dev/schemas/shopping/fulfillment.json, …
Spec and schema URLs let agents validate request/response shapes before sending. Broken URLs degrade runtime validation to "best effort" and cause strict agents to refuse the capability entirely.
What operations can agents actually perform?
Number of agent-readable capabilities under ucp.capabilities.
3 capabilities detected
Capabilities tell agents what operations to attempt. More coverage = more sessions complete end-to-end without falling back to scraping.
Store declares a checkout-namespaced capability.
dev.ucp.shopping.checkout detected
Without a checkout capability, agents can browse but can't complete a purchase end-to-end.
Service transports (REST/MCP/A2A/Embedded) declared on services.
REST, MCP
Transports tell agents how to actually call your endpoints. MCP and REST are the most widely supported today.
Tokenized payment handlers registered for in-conversation checkout.
Empty array: checkout-link redirect strategy
Tokenized payment lets agents complete purchases without redirecting the buyer to a hosted checkout. Empty by design is valid for WooCommerce-style native checkouts.
Live reachability check against https://www.wixapis.com/ecom/ucp/2d5231f9-7d9f-42a1-bc9c-907a5f33bc77.
Host responded with HTTP 404 — endpoint reachable but path may be method-restricted
A declared transport that does not respond means agents will hit a dead end the moment they try to use the capability — even if the manifest itself looks perfect.
Live reachability check against https://www.wixapis.com/ecom/ucp/2d5231f9-7d9f-42a1-bc9c-907a5f33bc77/mcp.
tools/list returned a JSON-RPC response in 167ms
A declared transport that does not respond means agents will hit a dead end the moment they try to use the capability — even if the manifest itself looks perfect.
The plain-text brief this store publishes to AI agents. Captured on the most recent check.
# GravelMan Series
> GravelMan est un événement sportif qui propose des défis de cyclisme d'ultra-endurance et de bikepacking sur des parcours mythiques, accessibles aux cyclistes de tous niveaux. L'événement offre une variété de formats, allant de 60 à 700 kilomètres, en version Gravel ou Route. Les participants sont assurés d'une aventure mêlant liberté, découverte et dépassement de soi, avec des cartes originales et des surprises à l'arrivée. GravelMan met l'accent sur l'autonomie et le respect de l'environnement, en encourageant l'utilisation des transports en commun et en minimisant l'impact écologique des épreuves.
## Homepage link
- [home](https://www.gravelmanseries.com/)
## Accès des agents IA
Ce site est proposé par Wix et prend en charge le protocole MCP
pour l'accès de l'IA agentive. Les agents IA peuvent se connecter directement pour récupérer le contenu du site en ligne et à jour, sans avoir besoin d'extraire les données.
- [**Point de terminaison MCP du site**](https://www.gravelmanseries.com/_api/mcp)
- [**Documentation MCP de Wix**](https://dev.wix.com/docs/develop-websites/articles/get-started/about-the-wix-site-mcp)
## Outils MCP disponibles
### GetBusinessDetails
Récupère les détails de l'entreprise et du site tels que le fuseau horaire, l'e-mail, le téléphone et l'adresse.
- Aucun paramètre n'est requis.
### SearchInSite
Recherche des informations sur le site.
- **searchTerm** - Le terme à rechercher dans le site.
### SearchSiteApiDocs
Récupère la documentation de l'API pour les solutions professionnelles Wix installées sur ce site et informe le client IA sur la manière d'utiliser les API. Utilisez cet outil pour interroger les articles et les services (au lieu de SearchInSite).
- **searchTerm** - Le terme à rechercher dans la documentation de l'API du site.
### GenerateVisitorToken
Crée une nouvelle session de visiteur et obtient un jeton d'accès visiteur pour le site.
Doit être appelé avant d'effectuer toute requête CallWixSiteAPI si aucun jeton de visiteur n'est déjà disponible dans le contexte.
- Aucun paramètre n'est requis.
### CallWixSiteAPI
Appelle les méthodes d'API sur le site pour effectuer des actions au nom d'un visiteur, telles que l'interrogation des données du site, la réservation d'un rendez-vous ou le commencement d'un achat.
- **visitorToken** - Jeton d'accès visiteur. Utilisez GenerateVisitorToken en premier si non disponible.
- **url** - URL absolue de la méthode d'API à appeler (ex. https://www.wixapis.com/...). Récupérez en utilisant SearchSiteApiDocs.
- **method** - Méthode HTTP à utiliser pour l'appel d'API.
- **body** - Corps de la requête sous forme de chaîne JSON valide.
### ReadFullDocsArticle
Récupère un article complet du portail de documentation pour développeurs Wix.
- **articleUrl** - URL d'un article dans le portail de documentation pour développeurs Wix.
### ReadFullDocsMethodSchema
Récupère le schéma complet d'une méthode d'API Wix. Doit être appelé avant d'appeler la méthode réelle via CallWixSiteAPI.
- **articleUrl** - URL de l'article de référence pour la méthode d'API Wix.
## Ce que les visiteurs peuvent faire via l'IA + le MCP
- Obtenir les détails de l'entreprise (informations de contact, emplacement, horaires)
- Découvrir les articles et services proposés
- Effectuer des réservations
- Démarrer un achat et être redirigés vers le site pour finaliser le paiement
- Poser des questions et obtenir des réponses pertinentes sans avoir à parcourir le site
## Remarques
- Aucune authentification n'est requise pour se connecter au point de terminaison MCP
- Seules les informations publiques accessibles sur le site sont disponibles
- Le contenu est toujours en ligne et à jour
- Pour recevoir des mises à jour des outils, implémentez un appel tools/list dès réception d'une notification de mise à jour d'outil| Cart ✓ Bound | 2026-08-25 |
| Checkout ✓ Bound | 2026-08-25 |
| Fulfillment Extension ✓ Bound | 2026-08-25 |
| Catalog Lookup ✓ Bound | 2026-08-25 |
| Catalog Search ✓ Bound | 2026-08-25 |
| UCP Status | Verified |
| Endpoint | https://gravelmanseries.com/.well-known/ucp |
| HTTP Status | 200 |
| UCP Version | 2026-08-25 |
| Transports | REST, MCP |
| Last Observed | |
| Last Full Index |
Drop the live grade into your README, status page, or marketing site. The badge auto-updates as your score changes.
Live preview[](https://ucpchecker.com/score/gravelmanseries.com)<a href="https://ucpchecker.com/score/gravelmanseries.com"><img src="https://ucpchecker.com/score/gravelmanseries.com/badge.svg" alt="UCP Score for gravelmanseries.com"></a>This score reflects data from the most recent crawl. Run a fresh check to update, or read the methodology to see how each signal is scored.
Track gravelmanseries.com's UCP score over time. We'll re-run the full check weekly and email you if the score drops, a capability regresses, or status flips.
Free. One email when something changes — no marketing. Privacy.
Track domains and get emailed when their UCP status changes.
By continuing you agree to our Terms & Privacy Policy.