Last updated: September 23, 2026
Dress2Ride collects the following data when you use the app:
Account information: Your account identity data from Google Sign-In or Sign in with Apple.
Closet data: Clothing items you add, including names, brands, categories, temperature ranges, photos, and generated images.
Ride data: If you connect Strava, we import ride history such as distance, duration, elevation, and timestamps to personalize recommendations.
Location data: When you plan a ride, your location is used to fetch weather forecasts. We do not use background location for the mobile app.
Import data: If you use Google Photos import, we process only the selected photos you choose to analyze.
Your data is used to provide personalized clothing recommendations, analyze clothing items from text, URLs, and photos, generate item imagery, and personalize recommendations from ride history. We do not sell your personal data. AI analysis is performed through secure backend services and your data is not used by us to train general-purpose models.
Your data is stored securely in Google Cloud Firestore under your user account. Each user's data is isolated and inaccessible to other users.
Each kit request also leaves a short technical log for 30 days: which steps ran, how long they took, and why a request failed. It holds no clothing, weather, location or text from the advisor, and it exists so we can fix failures you report by their reference number.
Firebase (Google): Authentication, data storage, cloud functions, and hosting.
Strava: Optional integration to import ride history. You can disconnect at any time.
OpenWeatherMap: Weather forecast data. Your location coordinates are sent to fetch forecasts and may be retained with ride plans and, if you consent to an enabled pilot, recommendation evaluation records.
Anthropic: AI analysis of clothing items and recommendations.
Google Gemini: Generated item imagery for closet items when you request it.
Google Photos Picker: Optional import of the specific photos you select. Dress2Ride cannot see the rest of your Google Photos library.
Direct photo uploads and ordinary Google Photos imports are analyzed to create editable closet drafts. Ordinary imports do not store the full source photo by default, but reduced review previews and reduced generation source assets may be retained with drafts so you can review the source evidence and request clean item images. If you choose to save an original image as an item photo, that stored file may retain metadata added by your device. When you add an item from a product link, the product image published on that page may be kept as a generation source asset for the same purpose, so the clean item image matches the real product.
For rider or kit photos, we do not request or analyze GPS/location metadata, face crops, race numbers, or background descriptions for ordinary closet capture. Benchmark test sets are an explicit opt-in quality workflow and are retained until you delete the benchmark test set or your account.
You can delete your account and associated data in the app from Settings > Delete Account. Account deletion removes closet items, AI drafts, ride and feedback data, generated images, user-scoped stored images, and benchmark test sets or benchmark runs you created. Primary account deletion starts immediately after confirmation, and some non-critical cleanup may complete asynchronously shortly afterward.
Public deletion instructions are available at /delete-account.
Participation in recommendation evaluation is optional. Capture is off by default. It requires an enabled pilot, a rider account the pilot admits (everyone with a verified account where enrollment is open, or an invited rider otherwise), and your explicit consent in Settings. Declining or withdrawing does not change your clothing advice.
If you join an enabled pilot, we retain the ride plan, route and location information supplied for the recommendation, weather, closet details, selected profile and past-experience information, provider requests and responses (including failed attempts), the final recommendation, and technical version and timing information. Your own recommendation grades are linked separately. These records help evaluate recommendation quality and reproducibility. Image references use item identifiers and storage paths rather than saved image access tokens.
Only the project owner may review identifiable pilot evidence, through an owner-only screen in the app that records every access; app-admin status grants no access. Experts and a second adjudicator may receive only separately approved blind/anonymized exports, never raw captures. Other riders cannot read your capture records. Joining this pilot does not authorize exporting your identifiable records as shared fixtures, sending them to additional AI providers, or training general-purpose models.
Pilot records are kept for 90 days after the recommendation, or 90 days after you last graded it, whichever is later. Incomplete captures expire 90 days after their first event unless you grade them later. Expert review does not extend retention. Cleanup runs hourly and normally removes expired records within a few hours of their deadline; during a backlog or outage removal can take longer, and the project owner is alerted when cleanup fails or stalls. Point-in-time recovery and scheduled backups are not enabled for this data, so deleted records cannot be restored after the database's one-hour internal version retention.
You can withdraw in Settings without deleting your account. Withdrawal blocks further capture and starts deletion of your pilot records; if cleanup is still pending, Settings lets you continue it. Account deletion also includes pilot records. These operations do not promise immediate physical erasure from backups. Limited consent-control information may remain after withdrawal to prevent delayed writes; it is removed as part of account deletion.
For questions about this privacy policy, contact us at support@dress2ride.cc.