Security
You stay in control. We stay in the audit trail.
Every message the AI drafts, every escalation it routes, and every filing that leaves your account is logged, timestamped, and attached to a named human — you.
Data protection and subprocessors.
Guest and reservation data is encrypted at rest and in transit. Access is scoped per account with the principle of least privilege — every AI call is narrowed, per call site, to the data that call needs. Guest-reply drafts (inbox + pilot demo) send the latest 5 turns of the guest thread and the property's operational house rules; nightly-rate suggestions send only address / city / bedrooms; the narrow chat relay is auth-gated streaming chat scoped per call site — a server-derived system prompt plus ≤ 20 turns of `{role, content}` messages, with payment, credentials, session tokens, payout details, uploaded IDs, and cross-tenant portfolio data excluded. Operational house rules (per property) are always sent when a property is in scope; your style block (tone + signature) is sent only when the personalization toggle is on. Drafts are sent directly from RentaraAI's server to the OpenAI API; property photos and permit PDFs are stored on Polsia R2; outbound guest messages are delivered via the Polsia email proxy; billing is handled by Stripe; sessions live in our Postgres database alongside your account row. The current subprocessor list with data categories, retention, and location lives on our subprocessor registry.
You are the approver.
Nothing leaves your account without your click. Drafts land in your inbox; you approve or edit; we send. Anything you mark off-limits the AI never sends — it escalates to the named human with the full context.
A ledger you own.
Every action is logged per property — who drafted, who approved, who sent, and when. The audit trail is yours to export, review, and share with your compliance team.
Want the operational detail?
The home page's control-and-security section walks through the exact permissions, escalation rules, and per-property guardrails.
Effective: 2026-09-11 · v2026-08-06