Security you can verify, not just trust
Your ledger is some of the most sensitive data you have. Here is exactly how we protect it — every claim on this page describes something real in the product today.
Role-based access
Owner, admin, and member roles per organization. Team members see and do only what their role allows — permission changes are logged.
Organization-scoped authorization
Every API request is checked against your organization. Your entities, transactions, and documents are never queryable from another customer's account.
Encrypted secrets storage
Connected-account credentials — like QuickBooks, Xero, and Google Drive OAuth tokens — are encrypted at rest with field-level encryption and support key rotation.
Audit trail on AI actions
Every AI categorization is recorded: what was coded, when, and with what confidence. Authentication and security events are logged immutably alongside it.
Stripe-hosted payments
Subscriptions run entirely through Stripe's hosted checkout and billing portal. Your card details go to Stripe — they never touch our servers or our database.
Daily AI spend caps
AI usage is metered per organization per day with hard caps, so a runaway job — or a compromised account — can't burn through unbounded processing.
Approval-gated AI
The AI can't silently rewrite your books. Uncertain categorizations wait in a review queue, adjusting entries land as drafts, and closed periods lock.
Encrypted in transit
All traffic between your browser and myTB — and between myTB and connected services — runs over TLS. Your data is yours: export it at any time.
Where we are on certifications
We don’t yet hold formal certifications like SOC 2 or ISO 27001, and we won’t put a badge on this page until an independent auditor has signed off on one. Plenty of startups do; we think that erodes the trust the badge is supposed to signal.
What we do instead is build the underlying controls first — immutable audit logging, role-based and organization-scoped access, encrypted credential storage, and approval gates on every AI action — so that a future audit certifies practices we already live by. If your firm has specific security requirements, ask us directly and we’ll give you specific answers.
Does your policy require firm-controlled hardware?
myTB On-Prem is a separate research product for firms evaluating local document storage, accounting data, and AI inference. It is not part of our cloud plans.
Found a vulnerability?
We value the security community and respond to every good-faith report. If you believe you’ve found a vulnerability in myTB, please tell us.
security@mytb.ai →