Privacy Policy

Your training data deserves precise language.

This policy explains how TriElite processes information in different app versions, when information may leave your device, and the limits of those protections.

Last updated: 14 October 2026

Live-release status: TriElite is publicly distributed through Apple’s App Store. This policy must remain aligned with the exact signed release build, production analytics and retention configuration, App Store privacy disclosure, public support process, and qualified legal review.

At a glance

Information stored on your device

You may provide race goals and dates, training availability and experience, planned sessions, completed activity information, Fuel entries, body and threshold inputs, Check-ins, Pain, Soreness, Sleep, Stress, Recovery context, coach preferences, notes, and privacy settings. The precise set depends on the features and permissions you use.

The current app is local-first: these records are designed to remain in app storage on your device unless a feature clearly requires a permitted system service or remote request. “Local-first” does not mean every operation is local, and it does not control device backups, Apple services, or information you choose to send elsewhere.

If you connect Apple Health, the app may request and process supported workout and health summaries. Apple controls Health authorization and does not expose every individual read-grant state to apps. Denied, unavailable, or missing information remains unknown rather than being treated as zero.

Location, Motion, Weather, and Apple Health

During an in-app workout, optional foreground Location can capture GPS coordinates, distance, elevation, and speed evidence. Optional background Location keeps that capture available when the screen is locked or the app is backgrounded. Motion & Fitness can help identify supported movement and moving-time evidence. These permissions are requested in context and remain separate; if they are denied or revoked, time-only recording remains available.

Captured route coordinates and derived workout evidence are designed to be stored with the local activity record. Temporary recovery checkpoints may also be stored on the device so an interrupted recording can be recovered, then cleared after the finalized activity is saved or the recording is explicitly discarded. Resetting or deleting local app data removes app-held copies, subject to device or platform backups.

If Apple Health workout export is enabled, a completed in-app workout summary, including supported distance and energy totals, may be written to Apple Health. The current app does not export its captured GPS route to Apple Health. Disconnecting Apple Health stops app synchronization and clears the app's pending Health work; records already held by Apple Health and native permissions remain managed in Apple's Health app and under Apple's policies.

Home may use Apple's WeatherKit only when foreground Location is already authorized. It uses recent coordinates to request current precipitation evidence and solar events for the decorative Home atmosphere; it does not request Location solely for weather, use background Location for weather, or send those coordinates through the plan or Coach model routes. Weather evidence is cached only for its bounded freshness period and otherwise falls back to a time-based atmosphere.

Planning, Coach, and on-device AI

In version 1.2.4 and later, Plan generation and Coach assistance run on your device using Apple Intelligence FoundationModels when available, with deterministic local behavior when it is not. Prompts and training context are not sent to a cloud AI provider.

AI can return a bounded plan proposal or user-facing wording. Deterministic code validates proposals, applies constraints, builds the concrete workout, and owns safety rules and calculations. You review proposed Plan changes before confirming them; Coach or Home guidance does not silently change the Plan.

Apple Intelligence availability depends on supported Apple hardware and software, supported system languages, regional availability, and device settings. If it is unavailable or not enabled, the app falls back to its built-in deterministic Coach and planning behavior.

Fuel photos, barcodes, and nutrition estimates

Fuel tracks carbohydrates, fluid, sodium, and calories. In version 1.2.4 and later, if you choose food-photo analysis, Apple Vision recognizes food and text locally and Apple Intelligence estimates nutrition from those observations on your device. The photo is not sent to a cloud AI provider. You review and can correct the estimate before saving.

When Apple Intelligence is unavailable, barcode lookup and manual entry remain available.

For a packaged-food barcode lookup, only the barcode digits are sent to Open Food Facts, not the captured image. Its community-supplied product and nutrition data may be incomplete or inaccurate, so verify it against the packaging before saving.

Versions 1.2.3 and earlier

Legacy AI services remain available for installed older versions. Standard planning may send minimized structured training, profile, race, availability and recovery context through our service to OpenRouter and a serving model provider. Coach may send your message, limited recent conversation and relevant training context. On-device-only planning avoids the plan server.

In older versions, optional food-photo analysis may send the selected photo through the disclosed remote service after your action. The one-photo “Analyze once” action does not change the processing mode for future requests. Manual entry and photo-free barcode lookup remain alternatives.

Strict zero-data-retention routing is disabled for these legacy AI services. Hosting, routing and model providers may process or retain requests under their own terms, settings, security practices and legal obligations. Our application service is designed not to intentionally store full prompts, photos or model responses in its application database; that does not guarantee no provider copies exist. Updating the app or deleting local records does not delete previously retained provider data. Do not enter emergency information, highly sensitive medical details or another person's information into free-text fields.

Analytics and diagnostics

App analytics uses a generated installation identifier and limited product-interaction, experiment, purchase, performance and diagnostic events to understand reliability and product use. You can turn off app analytics and session replay in Settings. The anonymous device identity and queued app events remain on your device so the same anonymous history can resume if you turn it back on.

From version 1.4.0, RevenueCat will send purchases, trials, renewals, cancellations and refunds, including product, price and currency, to PostHog to measure subscription revenue. This integration will be enabled after 1.4.0 is live and will operate whether or not app analytics is on. With app analytics on, these events can be linked to its anonymous identity. When app analytics is off, these events will use your RevenueCat customer ID and will not be linked to your app analytics identity or include your app activity.

Turning app analytics off does not delete events already received by PostHog. Those records remain subject to its configured retention and processing.

Apple and the App Store may separately process purchase, subscription, crash, and device information under Apple's policies and your device settings.

Website measurement

On public website pages, PostHog receives page views and clicks to the download page or App Store. This helps us understand which guides and search or AI referrals lead people to explore the app. These clicks do not establish an installation, purchase, or subscription.

Events contain the public page path, language, referring domain, acquisition category, and a random identifier for the browser-tab session. The session identifier expires after 30 minutes of inactivity and is kept in session storage, not a cookie. We do not send URL queries or fragments, referring-page paths, link text, form values, names, training information, or checkout handoff data in these events, and do not create a PostHog person profile.

Measurement respects the browser’s Global Privacy Control and Do Not Track signals. You can also turn off website measurement below; that choice is remembered in this browser for trielite.app. If you turned measurement off on 1406triathlontraining.com, set that choice again here; browser storage is separate for each domain. The website choice is separate from the app’s analytics and advertising settings. Previously received events remain subject to PostHog’s configured retention and processing.

Vercel Web Analytics also counts visits to public website pages without analytics cookies. It receives the public page URL with query strings and fragments removed, referring information permitted by your browser, and basic browser, device and country information. It follows the same website privacy choice, Global Privacy Control and Do Not Track signals. We do not send form contents, checkout data, training information or app identities to this website counter.

Advertising measurement

Starting in version 1.3.0, TriElite asks for Apple's App Tracking Transparency permission before advertising tracking begins. Only if you allow tracking and keep attribution sharing enabled in Settings can AppsFlyer receive install/session signals and the milestones onboarding completed, paywall viewed, and plan saved; TikTok can receive installation/session and these milestone signals, plus confirmed subscription trial/purchase event names without amounts. Meta can receive installation and app-open signals. RevenueCat can send subscription lifecycle events, purchase information, currency and revenue values, and matching identifiers to AppsFlyer and Meta. Advertising measurement links these signals with data from other companies' apps and websites. Matching identifiers can include the advertising ID (IDFA), vendor ID (IDFV), IP address, AppsFlyer ID, Meta anonymous installation ID and a hashed subscription customer ID. SDKs can also send device/software information and infer approximate location from IP addresses.

If you decline tracking, have not answered, or your device restricts it, these app tracking routes stay off and advertising matching identifiers are removed from RevenueCat. Apple SKAdNetwork/AdAttributionKit aggregate measurement does not require tracking permission. Training plans, Health values, Fuel entries, race details, names and free text are not included in advertising events. Turning attribution sharing off, or revoking permission in iOS Settings, stops new app tracking and queues removal of matching identifiers; withdrawal is synced when the app can reach RevenueCat. This does not delete events already received by providers or stop Apple and RevenueCat from processing your subscription. Purchases and app access work regardless of your choice. App analytics has a separate setting. From version 1.4.0, that setting will not stop the subscription revenue measurement described above. Older supported versions may have different attribution behaviour.

The dormant website-to-app purchase path also contains optional OpenAI advertising measurement. If that path is activated, the OpenAI Pixel is not loaded until you choose “Allow optional measurement.” With consent, it may receive page-view and checkout-started events plus OpenAI's opaque attribution references. After RevenueCat confirms an eligible first web purchase, the server may send a matching subscription-created event with the web product, price, and currency. These events use an opaque event ID for deduplication, set OpenAI's opt-out flag, and intentionally exclude names, email addresses, onboarding answers, race details, Health data, training data, RevenueCat customer identifiers, and payment details. Your choice is stored in this browser and can be reopened through “Ad choices.” The production activation gate requires OpenAI automatic advanced matching to remain disabled.

How information may be used and shared

Information may be processed to provide plans and workouts, record activity and Fuel, calculate Stats and Recovery context, perform requested AI or barcode features, maintain security and reliability, respond to support requests, process purchases, and comply with law.

Depending on the feature and production configuration, information may be processed by Apple and App Store services, hosting and network infrastructure such as Vercel, Open Food Facts for barcode lookups, RevenueCat for purchase and subscription services, Strava when connected, and analytics or attribution providers enabled in the signed build. In version 1.2.4 and later, AI plan, Coach, and food-photo processing runs on-device. Older supported versions can also use OpenRouter and selected model providers as described below. Each service processes data under its own terms, configuration, and legal obligations.

TriElite is not designed to sell personal information. Information may also be disclosed where required by law, to protect rights or safety, investigate misuse, or support a business transaction, subject to applicable law.

Retention, deletion, and access

App records stored locally remain until you delete or reset them, remove the app, or they are otherwise cleared by the device. Copies may remain in device or platform backups until those backups expire or are deleted under the relevant platform settings and policies.

A local deletion or app reset does not itself delete records held by Apple, legacy AI routing or model providers, hosting infrastructure, analytics or attribution services, RevenueCat, Strava, support systems, or other third parties. Their records follow their policies, configured retention, legal obligations, and applicable access or deletion request processes. We do not promise complete or immediate deletion from those providers.

Where applicable law grants access, correction, deletion, restriction, objection, or portability rights, requests need to use a verified public support channel and the applicable service-provider request processes. This page does not currently publish a verified privacy-contact address.

Children and teen athletes

TriElite is not intended for children under 13. Athletes aged 13 through 17 may be asked to attest that a parent or legal guardian has permitted their use. That checkbox is an unverified attestation; it is not identity verification or verified parental consent.

Age and guardian requirements vary by market and by race organizer. The 13+ rule, teen flow, under-13 reporting process, and guardian disclosures require ongoing qualified legal review for the markets where the app is offered. Athletes and guardians must also check the rules for their event.

Security and international processing

Reasonable technical and organizational safeguards are intended to protect information, but no device, transmission, or service can be guaranteed completely secure. Remote providers may process information in countries other than your own, where privacy laws may differ.

Changes and contact

This policy may change as the product and production configuration change. A revised policy will show a new last-updated date. Material changes should be presented as required by applicable law.

Apple’s App Store identifies RADIANT MANAGEMENT SYSTEMS PTE. LTD. as the app provider. This page does not currently publish a verified privacy-contact address; a public support and privacy-request process should be kept aligned across this page, the app, and the App Store listing.

Back to legal