What we collect, and what we deliberately do not.
A plain-language summary of LiftLine AI data handling, including the full list of outside services that receive anything and the things the product deliberately does not do.
We collect an unusually small amount, we do not sell or share any of it, and the list below of things we deliberately do not do is longer than the list of things we collect.
What we do not collect
Each of these was verified against the code, and several are things an app of this kind normally has.
- No analytics, telemetry, or product-usage SDK. None, in the app or on the website.
- No crash reporting and no session replay.
- No advertising identifier, no IDFA, no ad network, no ads.
- No fingerprinting, no cross-site or cross-app tracking, and no IP-based geolocation.
- No sale of data, no data brokers, no ad-tech recipients.
- No access to your microphone, camera roll, device camera, contacts, calendar, or health data.
- No background location, on any platform.
- We do not train any model on your data. The only model we train is a terrain classifier that learns from satellite and lidar imagery of mountains.
- The Snowcat iPhone app contains zero third-party SDKs of any kind.
Location works differently on iPhone and on the web
This is the biggest difference between our two platforms and it deserves to be stated rather than averaged.
- On iPhone
- Your location is used on the device to work out which resort you are at and how far you are from a lift. It is not sent to us.
- On the web
- Proximity to a lift is evaluated in your browser, and crossing a threshold produces an event that reaches our server. That is a difference from the iPhone behaviour and we are not going to describe the two as the same.
- Crowd reports
- A report you submit carries the lift, the value, and the time. It carries no coordinates: the endpoint accepts none and the table has no column for one.
Everyone else who sees anything
Derived from the code that makes each call.
| Service | What it receives | Why |
|---|---|---|
| Supabase | Your account, profile, preferences, crowd reports, favourites, entitlements, and authentication events. | Our database, authentication, storage, and server functions. |
| Apple | On iOS: your purchase and subscription. Your trips and passkeys live in your own private iCloud storage, which we cannot read. | Purchases, sync, passkeys. |
| OpenAI | Liftie: your conversation, your display name, and your Liftie memory if you enabled it. Separately, forecasting prompts containing no personal data. | The assistant, and crowd forecasting. |
| Anthropic | Liftie: as above. | The assistant. |
| Google (Gemini) | Liftie: as above. Separately, webcam frames for crowd counting, and forecasting prompts containing no personal data. | The assistant, crowd counting, forecasting. |
| DuckDuckGo | A short search query written by the model, sent with no name, account, or identifier of yours. | Letting Liftie look something up. |
| Resend | Your email address, your first name, and the contents of the message. | Sending account and service email. |
| Stripe (web payments only) | Your email address, your account ID as metadata, the amount, and the resort. | Web payments. |
| Apple and Google push services | If you turn on web notifications: the notification itself, which carries no name, email, or account identifier of yours. | Delivering a notification while our site is closed. |
| YouTube / Google | Your IP address and browser details when a webcam player loads, because the video streams from YouTube. | Live webcams. |
| AWS (Amplify) | Ordinary web server logs, including your IP address. | Web hosting. |
| Discord | Operational alerts from our collection service — crowd counts only, and no user identifiers of any kind. | Knowing when collection breaks. |
Webcams and the people in them
Our crowd counts come from public live webcams that resorts already broadcast. A frame is analysed for how many people are in a lift line and then replaced by the next capture. What we keep from a frame is a number, a confidence, and a timestamp — not the frame, and not an identity. We do not run face recognition and we do not attempt to identify anyone.
If you believe you can be identified in a webcam frame we process, write to privacy@liftlineai.app.
Your controls
You can sign out, clear what you have stored, and ask us to delete your account and its data. Requests go to privacy@liftlineai.app. California residents have specific rights, set out in the policy at https://liftlineai.app/privacy/.
What is changing
A rewritten privacy policy is in review. It was built by reading the code rather than editing the previous policy, and several clauses in it describe controls that are decided but not yet shipped. It will publish when those controls ship, not before — publishing a policy ahead of the code is how the current one came to overstate what we do.
- Privacy PolicyThe binding document.
- AI transparencyWhat Liftie sends, and to whom.
- Data sourcesEverything we read that is not yours.