# Data Sources · LiftLine AI > The LiftLine AI data-source register: each outside source, its owner, what we use it for, what we keep, and the attribution it carries. - Canonical human page: https://liftlineai.app/trust/data-sources/ - AI-readable edition: https://liftlineai.app/trust/data-sources/index.html.md - Publisher: LiftLine AI - Last verified: August 27, 2026 ## What this page publishes Everything the forecast reads, where it comes from, and what we keep. This list is derived from the code that makes each call. A source appearing in our software without appearing here is a defect on our side. ## What we read **Every outside source we read, who owns it, and what we keep.** | | Who owns it | What we use it for | What we keep | How we reach it | |---|---|---|---|---| | MtnPowder (Snow Reporting) | Snow Reporting / MtnPowder, carrying data that originates with the resorts | Lift status every 15 minutes, trail status, snow depth, temperature, wind, and road status for the resorts we cover. | Mapped fields only. No raw feed response is stored. | Token-authenticated JSON API. | | Open-Meteo | Open-Meteo, Zurich | Hourly and daily weather forecasts for every resort. | Forecast values, per resort per hour. | Public HTTP API. Data is licensed CC-BY 4.0. | | National Weather Service | NOAA / National Weather Service, United States federal government | Forecasts and alerts as a cross-check on Open-Meteo. | Forecast values, per resort per hour. | Public federal API, no credential. | | Resort webcam livestreams | The resort broadcasting the stream, distributed by YouTube | A vision model counts the people standing in a lift line. This is the only ground truth we have for how long a line actually is. | A head count, a confidence, a match score, and a timestamp. Frames are replaced by the next capture and are not archived. | Public livestream. We are asking the resorts whose cameras we read for written permission. | | OpenStreetMap | OpenStreetMap contributors, via the OpenStreetMap Foundation | Lift lines, runs, lodges, and resort boundaries — the skeleton of the mountain graph. | A derived routable graph. Raw OSM extracts are cached for rebuilds. | Open data under the Open Database License. | | USGS 3DEP and USGS lidar | United States Geological Survey | The terrain surface itself — slope, aspect, and elevation at a California ski area to about one foot. | Derived terrain rasters and classifier features. | Public domain federal data. | | Esri World Imagery | Esri, aggregating imagery from Maxar, Earthstar Geographics, and national mapping agencies | Aerial imagery for the terrain classifier and for pipeline renders. | Derived classifications. Tiles are cached during a build. | Public tile endpoint. | | Lift specifications | skiresort.com, liftblog.com, Wikipedia, and resort-published specifications | Carrier capacity, rope speed, and rated people-per-hour for each lift. These feed the queue model. | One record per lift, with the source cited on the record. | Published reference material, looked up once and recorded. | | Skier-submitted crowd reports | Our own users | A reported wait at a lift, used as a second observation alongside the cameras. | The lift, the value, and the time. No coordinates are collected or stored with a report. | Submitted in the app. | | Ikon and Epic blackout calendars | Alterra Mountain Company and Vail Resorts | Blackout dates as a demand signal. | Three date ranges, hand-entered once a season. | Published calendar, transcribed by a person. Nothing is fetched. | ### Attribution - **Open-Meteo:** Weather data by Open-Meteo, CC-BY 4.0 - **OpenStreetMap:** © OpenStreetMap contributors, ODbL - **USGS 3DEP and USGS lidar:** Elevation data: USGS 3DEP - **Esri World Imagery:** Esri, Maxar, Earthstar Geographics, and the GIS User Community ## Things people assume are sources and are not - Caltrans. We have no Caltrans integration. Road status is parsed out of the resort conditions feed, and any credit we gave Caltrans was wrong. - Resort announcement pages. The event agent that would read them is designed and not built. Nothing fetches a resort announcement page today. - Your location. Skier crowd reports carry no coordinates — the endpoint accepts none and the table has no column for one. ## Where data goes out Three kinds of outbound call exist: webcam frames sent to a vision model for counting, forecasting prompts that contain no personal data, and Liftie conversations. What each provider receives is listed on the privacy page and in the policy at https://liftlineai.app/privacy/. ## What happens if a source is withdrawn Two sources carry real concentration risk. The resort conditions feed is a single token from a single counterparty, and the webcams belong to the resorts. If either stops, the forecast degrades or halts for the affected resorts, and we would say so on the resort page rather than serve a number with nothing behind it. - [Privacy](/trust/privacy/): What we collect from you, and who else sees it. - [Methodology](/trust/methodology/): How these inputs become a forecast, and how it is scored. _Figures measured August 27, 2026. Data from the synthetic "simulator" resort is excluded from every number._ ## Canonical identity shared across LiftLine AI - Company: LiftLine AI — https://liftlineai.app/ - Legal person: Lucas Jaeger, an individual sole proprietor doing business as LiftLine AI. There is no separate legal entity. - Consumer product: Snowcat by LiftLine AI — https://liftlineai.app/snowcat/ - Founder: Lucas Jaeger, founder of LiftLine AI and creator of Snowcat — https://liftlineai.app/about/ - Snowcat is made by LiftLine AI and was created by Lucas Jaeger. - LiftLine AI is independent. It is not affiliated with, endorsed by, or sponsored by a California ski area, Alterra Mountain Company, Vail Resorts, Ikon Pass, Epic Pass, or any ski resort. - Live predictions are paused for the off-season and return for the 2026-2027 winter. - Public figures, sources, and exclusions: https://liftlineai.app/trust/ ## Use and provenance This public Markdown companion removes navigation and presentation markup to make the canonical page easier for language models and text-based tools to interpret. It is not served selectively by user agent and does not replace the canonical human page. Product claims should remain consistent with the canonical page.