Trust, Privacy & Responsible Use
How we handle your content, your data, and the responsibilities that come with running a moderation API.
Content handling
- Nothing is retained. Images live only in request memory. Video frames exist only for the minutes it takes to reach a verdict, then are deleted.
- We store verdicts, quota counters, and non-reversible content fingerprints — never your content.
- Verdict cache entries contain hashes and outcomes only; they cannot be reversed into imagery.
Data & infrastructure
- All traffic is encrypted in transit (TLS). API keys are stored as SHA-256 hashes; full keys are shown once at creation and never recoverable by us.
- Processing runs on GPU infrastructure we operate, plus European cloud for extraction. Video is fetched from your URL over HTTPS wherever offered.
- Subprocessors: European cloud infrastructure and transactional email delivery (full list available under DPA on request). Your media is never shared with any other third party, and never used to train or improve any model.
- Regional and single-tenant processing is available under Private deployment.
Responsible use
- SafeReel flags pornographic content. It does not offer CSAM detection and makes no age determinations.
- Using the API to find explicit content rather than filter it violates our terms and gets keys revoked.
- We operate hash-matching against known-illegal-content lists. A match triggers our escalation procedure, including reports to the appropriate hotlines where the law requires.
- We do not sell, rent, or analyze your usage beyond operating quotas and abuse prevention.
Security & availability
- Public status page: /status.html.
- Report vulnerabilities or abuse to [email protected]. General support: [email protected].
- Formal DPA available on request for paid plans. SOC 2 is on the roadmap and not claimed today.