1. Our privacy architecture
Visual Memory is designed around a local-first and minimum-access model. By default, full-resolution project media stay on your device. The service is intentionally limited so it does not hold a browsable archive of your originals.
Where data must leave the device (for cross-device search), we store only a lightweight semantic index — typically compressed thumbnails, voice transcripts you provide, and AI-derived labels — not your project library as a media vault.
2. Encryption & access limits
In transit: connections use HTTPS/TLS so credentials and data are encrypted between your client and our servers.
At rest (credentials): passwords are stored only as strong one-way hashes (bcrypt). We cannot recover your plaintext password; you can replace it via Forgot password on the sign-in page.
Cloud sync tokens: Google Drive / Dropbox access tokens are encrypted in your browser (IndexedDB + AES-GCM) and are never sent to Stippo servers. Full-resolution vault media stays on your device / your cloud folder (BYOS).
Originals by default: full-resolution files remain on the capturing device and sync to your Drive/Dropbox folder. Stippo servers do not store full-res vault media on the primary capture path.
Cloud index: the searchable index (understanding + optional thumbnails) is processed to provide the product. That index is account-scoped and not sold or shared for advertising.
3. What we process
- Account identity: name, email, authentication tokens.
- Billing metadata via Stripe (card details never stored by us).
- Memory metadata: titles, tags, transcripts, AI summaries, project links.
- Optional location if you enable GPS / EXIF capture.
- Thumbnails / embeddings needed for search (Hybrid Plan B).
4. AI processing
To make memories findable, submitted notes and (when provided) image thumbnails may be analyzed by automated vision/language systems. Outputs (titles, tags, summaries, embeddings) stay tied to your account for retrieval. Do not submit content you are not allowed to process under client or workplace rules.
5. What we do not do
- We do not sell your personal data.
- We do not use your project media for advertising profiles.
- We do not train public foundation models on your private library as a product feature.
- We do not access local unsynced originals on your device.
6. Your controls
- Export your account data (JSON) from Account.
- Delete individual memories at any time.
- Close your account with a three-step confirmation; we then delete account data, projects, memories, and associated media we hold, and cancel active subscriptions where possible.
7. Retention
We retain account and index data while your account is active. After verified account deletion, we remove associated personal data from production systems except records we must keep for legal, tax, or fraud prevention (e.g. Stripe invoices).
8. Contact
Privacy requests: use the in-app Account tools first (export / delete). For additional inquiries, contact the operator of your Visual Memory deployment via the email published on the service homepage or billing receipts.