Security
We handle ID documents and selfies, so the design starts from what could go wrong. This is how the service is protected, in plain words. The binding description is Annex II of the data processing agreement.
Encryption
- All traffic uses TLS. The service sends HSTS and a strict content security policy.
- Every image and video is encrypted by the application with its own key before it is written to disk. The keys are kept apart from the database. Deleting a key makes the file unrecoverable.
- Names, addresses, emails and review notes are encrypted in the database. For document numbers we keep only the last four characters and a keyed hash.
Who can see what
- Every client sees only its own data. Isolation is enforced on every query.
- Evidence is never served as a public file. Each view goes through an authorisation check, is not cached and is written to an audit log.
- Reviewers see only the submissions they have claimed, must use two-factor authentication, sign a confidentiality agreement and cannot review their own organisation.
- The platform owner does not browse evidence by default. Emergency access needs a typed justification, lasts 15 minutes and raises an alert.
Audit trail
- Decisions, evidence views, exports, erasures, settings changes and staff changes are logged without applicant personal data.
- Logs, error reports and chat alerts never contain applicant data, tokens or media.
Availability and backups
- The service runs on dedicated servers in the EU, with the media on a separate volume and an alert before the disk fills.
- Backups are encrypted and verified every week, and an alert fires if one goes stale.
Safe by default
- Uploads are re-encoded and size-limited, and metadata such as EXIF is removed.
- Webhooks are sent only to public HTTPS addresses and are signed so you can verify them.
- Dependencies are audited on every build.
Incidents
- We have a written breach runbook. If a personal data breach affects you, we notify you within 24 hours so you can meet your own 72-hour duty.
LegitFlash is not for AML/KYC-regulated onboarding and does not provide qualified or legally binding digital identities.
Found a vulnerability? Write to support@example.com. We read every report.