Security & Data Handling

Last updated: August 2026

You upload business plans, financials and sometimes unpublished IP. This page explains, plainly, what happens to that material. Where a statement depends on a contract or a setting, we say so.

The short version

Your content is never used to train AI models — not by us, not by Anthropic.
Encrypted end to end: TLS 1.2/1.3 in transit, AES-256 at rest.
Bounded retention at the AI provider. Anthropic deletes API inputs and outputs from its systems within 30 days and does not use them to train its models.
You control deletion. Delete any session at any time; it is removed immediately and permanently.

How the AI provider sees your data

An AI model has to read your text to analyse it — that is true of every provider, and no vendor can honestly claim otherwise. What can be controlled is what happens around that moment, and that is what we control: the connection is encrypted, Anthropic processes requests under commercial terms that prohibit training on customer content, and retention is bounded. Anthropic deletes API inputs and outputs from its systems within 30 days and does not use them to train its models. Anthropic's trust and safety systems may retain content that is flagged for policy violations, as disclosed in their data-retention documentation.

How long we keep your session

During the closed beta, with your agreement, your sessions and reports are retained so we can learn from them and improve the product. You can delete any session at any time; deletion is immediate and permanent.

Deleting a session removes the report, the conversation, the extracted text of your uploads, and any cached PDF or diagram renders. What survives is deliberately content-free: the payment record, token counts for billing, and the record that you acknowledged the download terms.

Controls

Area	What we do
In transit	TLS 1.2 and 1.3 only, modern AEAD cipher suites, HSTS. Older protocols (SSL, TLS 1.0/1.1) and weak ciphers (RC4, DES, 3DES, MD5) are disabled.
At rest	AES-256 disk encryption on the database (Supabase) and at the AI provider (Anthropic). Certificates use RSA-2048 or ECDSA P-256 keys.
Access	Every record is tied to your account and tenant; database row-level security plus application checks prevent one customer from reading another’s data. Invite-only accounts; no public sign-up.
Uploads	Documents are parsed in an isolated, resource-limited process; only the extracted text is kept with your session. The raw file is deleted immediately after extraction.
AI provider	Anthropic’s commercial API. Requests go server-to-server over TLS; no browser ever talks to the AI provider directly, and no user-supplied API keys are accepted.
Analytics	Product analytics record which steps you reach and how long they take — never document text, answers, or report content (enforced by an allow-list of event fields).
Payments	Stripe. Card details go straight to Stripe; we store only the payment reference and amount.

Sub-processors

Anthropic (AI analysis; no training on customer content; SOC 2 Type II) · Supabase (PostgreSQL database and authentication; SOC 2 Type II; AES-256 at rest) · Stripe (payments; PCI DSS Level 1) · Resend (transactional email).

What we don't do

We don't train models on your content, and our provider doesn't either.
We don't sell or share your content with other customers or advertisers.
We don't log the text of prompts, documents or reports in application or analytics logs.
We don't hold ourselves out as certified where we aren't: nDirector itself is not yet SOC 2 audited; our infrastructure providers are.

Reporting a security issue

Email security@ndirectorsuite.com. We acknowledge reports within two business days.

Related: Privacy Policy · Terms of Service