Trust Center

Security built for supply chains.

Your supplier list, product catalog, purchase orders, and shipment data are some of your most sensitive operational assets. We built CartonFlow assuming a competitor could be on the other side of the wall.

🔐

Privacy Mode

Optional client-side encryption. Your supplier and product names live encrypted on our servers — even CartonFlow staff can't decrypt them.

🛡️

Tenant isolation

Every database read filters on your account ID at the query layer. Account A can't see Account B's data even in error scenarios.

📋

Full audit trail

Every action by you, your team, and our AI assistant is logged to an append-only audit table. Investigate any change later.

Differentiator

Privacy Mode: data we
literally cannot read.

For tenants that turn it on, product names, ASINs, SKUs, and supplier names are encrypted in your browser before they reach our servers. The decryption key never leaves your device.

If we got subpoenaed, hacked, or had a rogue employee — there's nothing useful for them to read. The data on our database is opaque ciphertext.

Read the technical details in our Privacy Policy →

🔒

Your data, encrypted at every step.

Standard encryption everywhere, with stronger protection available for tenants who need it.

TLS 1.3 in transit

Every byte between you and CartonFlow flows over TLS 1.3. HSTS preload directive tells browsers to never connect over HTTP again.

AES-256 at rest

Database storage is AES-256 encrypted at rest by default — the standard the U.S. government uses for top-secret data.

Argon2id passwords

Passwords are hashed with Argon2id — the OWASP-recommended algorithm. Even if our database leaked, passwords stay protected.

Two-factor authentication

TOTP-based 2FA for any account that wants it. Required setting available for accounts with elevated permissions.

One database, but every tenant lives alone.

Multi-tenant architecture done right: isolation enforced at the lowest possible layer, not after-the-fact in application code.

Query-layer tenant filter

Every read filters on your tenant ID. Even an internal bug can't accidentally show one customer's data to another.

JWT session tokens

Short-lived signed tokens carry your authentication. Revocable instantly via password change or 2FA reset.

Role-based permissions

Owner, distributor, 3PL partner, and read-only roles — each only sees what their job requires.

Production access logged

Internal access to production infrastructure is gated, MFA-enforced, and recorded in an immutable audit log.

Will CartonFlow get my Amazon account flagged?

Short answer: no. Here's why we're confident saying that.

We use Amazon's official SP-API the way Amazon intends.

CartonFlow integrates with Amazon Selling Partner API through Amazon's standard OAuth flow. You explicitly authorize the connection from your Seller Central account; you can revoke it anytime.

  • OAuth-scoped, revocable: we never see your Amazon password. You grant only the permissions we ask for, and you can revoke from Seller Central with one click.
  • Rate-limit compliant: we respect Amazon's per-endpoint rate limits with exponential backoff. You won't get throttled because of CartonFlow.
  • Read-mostly access: 95% of our calls are read operations (sales velocity, FBA inventory, shipment status). Reads cannot flag accounts.
  • No policy-violation surfaces: account flagging happens for review manipulation, counterfeits, policy violations, suspicious login patterns. None of those are things CartonFlow touches.

The companies that touch your data.

Every B2B service relies on vendors. We pick ones we trust, with regions and purposes clearly defined. Updated as our stack evolves.

Vendor
Purpose
Region
Fly.io
Application hosting (cartonflow.io app servers)
US
Supabase
Postgres database hosting (encrypted at rest)
US
Cloudflare
DNS, DDoS protection, edge caching
Global edge
Stripe
Payment processing (PCI DSS Level 1)
US
Resend
Transactional email delivery
US
Sentry
Error monitoring (no PII in error payloads)
US
Upstash
Redis-based rate limiting
US
OpenAI / Anthropic / Google AI
AI assistant inference (zero-retention agreements; data not used to train models)
US

We notify customers via email and update this page at least 30 days before adding a new subprocessor. View the full subprocessor list →

What happens when something goes wrong.

Real security is the process behind the product. Here's how we operate.

01

Incident response

Documented runbook for security incidents. Affected customers notified within 72 hours per industry standard. Post-incident report published within 7 days.

02

Vulnerability management

Dependencies scanned weekly via Dependabot. Critical CVEs patched within 7 days; high-severity within 30. Reviewed and tracked in a public-style log.

03

Backup & restore

Daily automated database backups. Restore drills run quarterly to verify backups are usable. Point-in-time recovery available for the last 7 days.

04

Responsible disclosure

Found a vulnerability? Report it safely to [email protected]. We respond within 2 business days and won't pursue legal action against good-faith research. Read the full disclosure policy →

05

Audit logs

Every action that changes your data is recorded — who did it, when, from what IP, what was the before/after. Available to account owners on request.

06

Data Processing Agreement

DPA available on request for any customer needing one for GDPR, CCPA, or internal procurement. Email [email protected].

Compliance status, plainly stated.

We'd rather tell you exactly where we are than imply we have certifications we don't. Here's the honest picture.

GDPR-aligned

Data subject rights honored: access, correction, export, deletion. EU customers supported with SCCs in our DPA.

CCPA aligned

California consumer privacy rights honored. We don't sell your data and never have.

SOC 2 — on the roadmap

We're not yet SOC 2 certified. If your procurement requires it, talk to us — we'll share our timeline and the controls we've already implemented from the framework.

Pentest — on the roadmap

No third-party pen test report yet. Enterprise customers can request our internal security review and threat model under NDA.

Have a security question we didn't answer?

Or need a DPA, security questionnaire response, or pen-test report under NDA?

© 2026 CartonFlow by Pevara Systems