MFLab > Memora

Privacy Policy for Memora (Re-Moment)

Effective Date: September 1, 2026

This Privacy Policy explains how Memora (“we”, “our”, or “us”) handles information when you use the Memora mobile application (the “App”). Memora is designed as a local-first photo indexing and memory-story rendering app.

1. Photos Access and Local Processing

Memora requests Photos library access so it can index assets, build recommendations, prepare story drafts, and render videos. The App may process photo metadata, thumbnails, image content, face-count signals, scene signals, quality signals, captions, and location/date information that is available through Apple’s Photos APIs.

Standard indexing, recommendations, search, story review, and video rendering run on the user’s device. Content-bearing app data such as indexes, generated stories, render jobs, preferences, checkpoints, diagnostics, and person profiles is stored locally on your device in the App sandbox with SwiftData and related local storage. Usage statistics are handled as described in Section 4. We do not have direct access to your local Photos library.

1.1 On-Device Photo Analysis and Face Data

Memora utilizes Apple’s local Vision framework directly on your device to detect face features, face counts, and face bounding box coordinates. We are committed to complete transparency regarding facial analysis data. Below are detailed answers regarding how face data is handled:

2. Pro Vision AI

Pro Vision AI is optional. If the user explicitly opts in and enables it, downscaled photo previews and related text context (such as scene tags, face count, capture time, and city/country) are transmitted securely to Memora Cloud and our configured third-party AI service provider, Ollama Cloud API, for caption generation. Original photo files and Exif metadata are never sent.

The service immediately deletes all image bytes after processing and does not retain them for product use, model training, or dataset creation. Operational records may be retained, including sanitized request metadata, generated caption text or provider response text, request identifiers, request status, quota counters, entitlement identifiers, device hashes, timestamps, and error details. These records are used for entitlement validation, quota enforcement, idempotency, debugging, security, fraud prevention, and abuse prevention.

We and Ollama Cloud API do not use the user’s photos or generated private photo metadata to train or fine-tune AI models.

3. Purchases, Entitlements, and Quota

In-app purchases are processed by Apple. We do not receive credit card or other payment details. To validate purchases and manage Pro Vision AI quota, the App and backend may process StoreKit transaction claims, product identifiers, entitlement status, quota usage, server-issued access tokens, and device hash values.

4. Usage Statistics

Starting with app version 1.1.5, usage-statistics sharing is on by default. If your device region is in the EU, the EEA, or the United Kingdom, sharing is instead off by default and is transmitted only if you turn it on, in line with local consent rules. You can turn sharing off (or, in those regions, on) at any time in the App’s Activity tab, under Usage Statistics, and turning it off stops all future transmission and requests deletion of the data associated with your installation, as described below. In versions before 1.1.5, sharing was off by default and was transmitted only if you turned it on; if you had not yet made a choice, the App kept one-time funnel milestones only on your device.

When usage-statistics sharing is on (the default from version 1.1.5), the App sends:

These records are used to measure the product funnel and aggregate activity, improve service quality and features, reconcile uploads, prevent abuse, and publish rounded aggregate adoption or usage figures. They are not used for cross-app tracking or targeted advertising. Usage-statistics payloads never include photos or videos, captions, locations, face counts or coordinates, person clusters, search terms, story titles, or photo-library identifiers.

A hashed installation identifier is a pseudonymous identifier, not a guarantee of complete anonymity. Because records for the same installation can be associated over time and may also be linked to a Memora backend account when available, we treat this usage information as linkable data. From version 1.1.5, we share it by default to improve the App, and we give you an always-available control to opt out in the App’s Activity tab, under Usage Statistics; opting out stops future transmission and requests deletion of your installation’s data as described below. We follow the transparency principles in the Personal Information Protection Commission’s privacy-policy guidance.

Raw usage increment events are retained for up to 365 days. Daily activity records are retained for up to 1,095 days. The latest per-installation snapshot and bounded cumulative totals are retained while sharing remains enabled or until the applicable account or analytics-deletion process removes them. Turning sharing off stops future transmission, clears locally queued analytics, and requests deletion of the server-side usage snapshot, raw events, daily activity, and cumulative totals associated with that installation. If the device is temporarily offline, the App retains the deletion request and retries it when a backend connection is available. You may also contact us using Section 10 if deletion does not complete.

5. Diagnostics

Memora may use Apple-provided diagnostics such as MetricKit to help identify crashes, hangs, launch issues, CPU exceptions, and disk-write issues. These diagnostics are intended for stability troubleshooting and do not include the user’s original photo files.

6. Support Requests

Support requests are optional and user-initiated. Nothing described in this section is collected unless you choose to contact us.

6.1 In-App Support Center

When you submit a request from Settings > Customer Support, the App sends:

The marker is a keyed hash computed on our backend; the installation identifier itself is never transmitted to the support system. It exists so the App can show you your own requests and enforce the limit of three open requests per device, and it cannot be used to identify you outside the support system.

Requests are stored as issues, and attachments as files, in a private repository hosted by GitHub, Inc. (a Microsoft company) in the United States. The repository is not public and is accessible only to the Memora maintainer. Your device never contacts GitHub directly: requests are proxied by the Memora backend, which scopes every read and write to the device that created the request, so one device cannot read another device’s request or attachments.

Screenshots are selected by you and are transmitted as you provide them. Please avoid attaching images showing information you do not wish to share, such as other people’s photos, private messages, or account details.

Nothing else from the App is included. Support requests never contain your photo library contents, indexes, captions, locations, face data, person clusters, search terms, or Apple ID.

We use the information in a request to reproduce and resolve the issue you reported and to reply to you. Requests and their conversations are retained until deleted. You may ask us to delete a request at any time through the contact channel in Section 10.

6.2 Email and Web Contact

If in-app support is unavailable, the App offers to open your own email client; in that case the App itself transmits nothing. The support form on our website sends the email address, subject, and message you enter, together with your browser user agent and network IP address, through our email delivery provider, Resend, Inc., to our support inbox.

6.3 Rate Limiting

Our backend processes the network IP address of support requests transiently to enforce rate limits and prevent abuse. It is not stored with your request and is not transmitted to GitHub.

7. Security and Retention

We use reasonable technical and organizational safeguards appropriate for a small application service, including Apple platform sandboxing, Keychain-backed app secrets where applicable, server-side entitlement checks, short-lived server tokens, request size validation, idempotency checks, and quota enforcement.

Operational backend records are retained only as long as reasonably needed for the purposes described above, unless a longer period is required for legal, security, accounting, or dispute-resolution reasons.

8. User Choices

Users can control Photos access in iOS or iPadOS Settings and choose whether to use Pro Vision AI. Usage-statistics sharing is on by default from version 1.1.5 (and off by default in the EU, the EEA, and the United Kingdom), and you can turn it on or off at any time in the App’s Activity tab, under Usage Statistics; earlier versions required you to turn it on everywhere. Disabling Photos access or Pro Vision AI may limit some App features. Usage-statistics sharing does not affect core App functionality. Submitting a support request is always a deliberate action, and you choose which details and screenshots, if any, to include.

9. Changes to This Policy

We may update this Privacy Policy from time to time. Material changes will be reflected by updating the Effective Date and publishing the revised policy on this page.

10. Contact

For privacy questions, including requests to delete a support request, contact MFLab-AI (Sang Hun Kim) through Settings > Customer Support in the App, the support form on our website, or the contact channel provided with the App.