nausika

légal · conditions

Conditions.

Dernière mise à jour

Beta

Nausika is in public beta. Availability, features, and accuracy may change. We do our best, we do not guarantee uptime.

Not a substitute for responsible navigation

Service availability & jurisdictions

Nausika is hosted in the European Union (Amsterdam region) by Railway, with edge caching and image storage on Cloudflare. Connected user data resides in the EU; see Privacy for the full subprocessor list and data-residency statement.

The service is offered to users in jurisdictions where Anthropic's Supported Countries policy permits the use of Claude. We do not block access at the network level, but use of Nausika via Anthropic-operated AI clients is subject to that list. Connections via other MCP-compatible AI clients are subject to those clients' own jurisdictional policies.

Data contributions

When your AI proposes an edit through Nausika, that edit is reviewed and, if approved, merged into a shared atlas that benefits every user. You grant us the right to use that contribution under a permissive license.

Data sourcing & web crawling

Place data — marinas, anchorages, fuel stations, lighthouses, services — comes from three classes of sources, in this order of preference:

  • Open-licensed datasets. Primarily OpenStreetMap (ODbL 1.0) and Open-Meteo (CC BY 4.0). See Credits for the full list and attribution.
  • Factual information from public web pages — names, coordinates, declared services, declared berth count, public contact details — extracted from the publicly accessible pages of marina operators and tourism portals.
  • Community contributions reviewed by a human curator before they land in the public dataset.

Web crawling for the second class is offline and batched, never live: when you invoke a Nausika tool, no request is made to a third-party marina website. Tools serve from our own database. Our crawl policy:

  • User-Agent nausika-marina-enrichment/1.0 (+https://nausika.app), always identifiable and contactable.
  • robots.txt directives from source sites are respected.
  • Each marina is re-crawled at most once every 30 days.
  • Only factual data is extracted: name, coordinates, declared services, declared berth count, public contact details. We do not extract user-generated content, reviews, photos, or copyrighted marketing copy.
  • Every record carries provenance: the upstream source URL is stored alongside the attribute and can be exposed in tool responses (the sources field of nausika_search_places).

Nausika's curated coverage is the Mediterranean basin, where most source portals are EU-based and may be protected by the European Database Directive (96/9/EC) sui generis right. The crawling policy above is designed to extract only insubstantial portions of any single source database, on a non-systematic basis, in line with normal use of those databases.

Take-down

If you operate a marina, a portal, or any source whose data appears in Nausika and you want it removed, write to [email protected] with a reasonable reference to your property. We aim to action take-down requests within 7 days. Removed records are tombstoned (we keep a record of the take-down so the data does not silently come back from a future re-crawl).

Contact

Email [email protected].