Trust
Verification, provenance, and ARD trust.
Snowdata combines resort ground truth with 7 weather models and 12M+ historical records. This page documents how agents and discovery services should evaluate our data and catalog entries.
Verification methodology
Every snow report in the Snowdata network originates from a verified resort or licensed feed — not scraped aggregator pages. Resorts submit through eight channels (email, SMS, photo, API, scrape, public feeds, AI assistant, webhook). Snowdata reconciles submissions against seven NWP weather models hourly and surfaces disagreements for review.
Live MCP responses and Answer Engine answers include attribution, timestamps, and evidence packs pointing to the underlying resort or model sources. Intelligence cards on Snowdata and SnowSure label data readiness and never expose internal ops methodology in public copy.
ARD trust manifest
Our ai-catalog.json entries include a trustManifest binding the publisher identity to https://www.snowdata.ai with attestations for data provenance (this page) and the Open Source Snow Spec.
- Identity: HTTPS FQDN
https://www.snowdata.ai - Provenance: publishedFrom Snowdata Inc.
- Federation: linked SnowSure and mcp.snowdata.ai sub-catalogs
Discovery surfaces
- ARD ai-catalog manifest: https://www.snowdata.ai/.well-known/ai-catalog.json
- OpenAPI — https://www.snowdata.ai/.well-known/openapi.json
- ARD search registry — POST https://www.snowdata.ai/api/ard/search
- For AI systems hub
Enterprise compliance attestations (SOC2, GDPR) will be linked here as they become available. Security questions: info@snowdata.ai.