Security you can verify
We build privacy infrastructure, so our own posture has to hold up. Here's how the platform protects data, and an honest account of what we do and don't claim.
How data is protected
Tenant isolation
Multi-tenant Postgres with row-level security keyed per request. A forgotten scope returns zero rows, not another tenant's data. Proven against real RLS.
Crypto-shred erasure
§12 erasure destroys the per-principal key. Sealed PII is unrecoverable, including from WAL/backups. This is not a soft delete.
Non-repudiation
Per-tenant signing (HMAC / ECDSA / KMS). KMS-backed keys keep the private half out of the process and survive restarts.
Tamper-evident ledger
Append-only hash chain. Any tampering breaks verification, which anyone can run with the published key.
Edge security headers
Served from Cloudflare with CSP, HSTS, X-Frame-Options and a strict permissions policy.
Least-privilege keys
Publishable / secret / read-only / operator key classes; secrets never reach the browser; per-IP rate limiting.
What we claim, and what we don't
We do claim
- Evidence-grade, tamper-evident consent (§6/§6(10)).
- Real crypto-shred erasure and RLS tenant isolation.
- A fail-closed §9(3) children gate.
- Self-host so your data can stay in your VPC / region.
We do NOT claim
- To be a registered §6(10) Consent Manager (that registry opens later in 2026).
- "DPDP-certified" or "Board-approved". No such certification exists to hold.
- Legal advice. Notice content and DPDP guidance are counsel-gated.
- SOC 2 / ISO 27001 yet. It's on the roadmap; we'll publish reports when audited.
Verify it yourself
Run the ledger verification, inspect the RLS isolation, and see crypto-shred erasure in a live demo.