Privacy Policy

Short version: Scenemark Stories is local-first. It analyzes only the photos and videos you choose, keeps story generation on device, and never names people automatically. A name enters a story only when you type it, choose it from Contacts, or assign it to a local face group yourself.

What Scenemark Stories collects

Selected photos and videos

Scenemark Stories uses Apple's photo picker and PhotoKit to work with the photos and videos you select. The app may load thumbnails, previews, selected image bytes, selected video frames, dates, dimensions, and available location metadata so it can build a grounded memory. It does not scan your full photo library.

Media analysis

The app uses on-device Apple frameworks such as Vision, Core ML, Natural Language, Speech, Sound Analysis, and VisionKit to prepare bounded observations from selected media and user-provided text. Raw face crops, embeddings, full audio, raw transcripts, raw OCR dumps, PhotoKit identifiers, confidence scores, dimensions, face counts, diagnostics, and system instructions are not used as public story text.

People and Contacts

The app may show local face-group suggestions, but those groups are not identities. You can add a person by typing a name, choosing a display name from the system Contacts picker, or assigning a typed or contact-selected name to a local face group. Contacts are user-selected only; Scenemark Stories does not silently match faces to Contacts or use the Photos People database as an identity source.

Location context

When available and permitted, location context may be used to help ground a memory with the right place and time. You can control location access in Settings > Privacy & Security > Location Services > Scenemark Stories.

Notes and story details

Titles, notes, dates, memory type, story style, aesthetic cue, confirmed people, and saved scene details are stored so the app can create, show, search, export, and back up your memories.

Local model files and diagnostics

If you import or use a local Gemma/LiteRT model file, Scenemark Stories stores that model in the app's local Documents area and marks model files to be excluded from iCloud backup where the app controls the file. Operational diagnostics for local model setup and generation may be stored in the app container to help troubleshoot on-device behavior.

What stays on device

• Selected-media analysis is designed to run locally.
• Story writing uses the local Gemma/LiteRT runtime.
• Raw media analysis is kept separate from visible story prose.
• People are never named without your explicit action.
• No advertising identifier is used.

Exports and backups

Public text export is intentionally narrow: it includes story-safe prose, summary, and confirmed people, and excludes source notes, PhotoKit identifiers, raw OCR, Vision summaries, diagnostics, and backup-only fields. Backup export is a sensitive preservation format and may include data needed to rebuild your local library, such as selected media identifiers, coordinates, summary signals, and memory metadata.

What we don't do

• We don't sell or rent your data.
• We don't track you across apps or websites.
• We don't use the IDFA or advertising SDKs.
• We don't infer names, relationships, or identities.
• We don't upload your selected media to servers we operate.

Apple services and crash reports

Scenemark Stories relies on Apple system frameworks and system pickers for media, Contacts, local analysis, files, and app operation. Standard Apple crash reports may be made available to us through TestFlight or App Store Connect if you have opted in to share crash analytics with developers in iOS Settings.

Children

Scenemark Stories is not directed at children under 13. We do not knowingly collect personal information from children under 13.

Changes to this policy

If this policy changes, we'll update the “Last updated” date at the top of this page.

Contact

Questions about this policy or how Scenemark Stories handles data: basiliosp@outlook.com