Skip to main content
Tradera.AI

Security & privacy

Your plans don't leak.

Builders trust us with commercially sensitive project information. We protect it by design — not as an afterthought. The full security model is documented in our public SECURITY.md.

Australian data residency

Customer data — including uploaded plans — lives in Supabase's ap-southeast-2 (Sydney) region. APP-compliant by default. NZ Privacy Act 2020 compatible.

Encryption in transit & at rest

TLS 1.3 for everything on the wire. AES-256 at rest for the database and storage. PDF uploads in private buckets, accessible only through short-lived signed URLs.

Tenant isolation by database

Postgres Row-Level Security enforces "your org sees only your org's data" at the database layer — not just in app code. Every policy is automated-tested.

No AI training on your plans

Tradera uses the Anthropic Claude API. Anthropic's policy is that API content is not used to train their models. We don't pass your plans through any service that would.

Audit log + observability

Every plan upload, AI call, RFQ draft, and billing event is recorded in an append-only audit log. Errors go to Sentry with PII scrubbed before transmission.

Breach notification within 72 hours

In the event of a notifiable data breach, affected customers are notified within 72 hours — matching the Australian Notifiable Data Breaches scheme.

Sub-processors we use.

These are the third parties that touch any part of customer data, what they do, and where.

ProviderLocationPurpose
AnthropicUSAI inference (no training on API data).
SupabaseAustralia (Sydney)Database, storage, authentication.
VercelEdge — SydneyWeb hosting and edge runtime.
StripeGlobal / PCI-DSS L1Payment processing.
InngestUSBackground-job orchestration (no plan content stored).
ResendUSTransactional email.
SentryEU/USError monitoring (PII scrubbed).
PostHogEU/USProduct analytics (autocapture disabled).
Report a vulnerability: security@tradera.ai · 48-hour acknowledgement SLA.