Security

This page summarises how Dolemai protects your information. Full detail is in our Privacy Policy.

How documents are handled

Dolemai reads the documents in the folders you connect, extracts the project intelligence, and discards them. Documents are never written to disk on our servers and are never saved to our database. Only the structured registers and reports are kept.

Reading those documents requires sending their content to Anthropic's Claude API. Anthropic retains that content for up to 30 days for safety and security purposes, then deletes it, and under their commercial terms it cannot be used to train AI models.

We think that is worth stating plainly rather than describing our own servers and leaving the rest unsaid.

Where your data lives

Your account and project data is held in the United Kingdom (London). Payments are handled by Stripe; we never see or store your card details.

Protecting your connections

The tokens that let Dolemai read your Microsoft 365 folders are encrypted with AES-256-GCM by our application before storage, on top of encryption at rest by our database provider. They are used only for the folders and mail channel you have connected, and you can revoke access at any time from Project Settings or from your Microsoft account.

Access controls

Row-level security is enforced on every table, so your data is reachable only by your account. Multi-factor authentication is available using an authenticator app; if you sign in with Microsoft, your organisation's own multi-factor policy applies.

Access to customer data is restricted to authorised personnel, only where necessary to provide support, investigate a fault or maintain security.

No tracking

Dolemai uses no third-party analytics, advertising or tracking services, and no error-reporting service that could capture your content. The only cookies we set keep you signed in.

AI outputs

Everything Dolemai produces is a draft for professional review. See our Terms for what that means in practice.

Responsible disclosure

If you believe you have found a security vulnerability, contact [email protected]. Please include enough detail for us to reproduce the issue and give us reasonable time to respond before any public disclosure. We will work with you in good faith.