Built for trust from the first email.
Tenant isolation
Every record is scoped to a management company and community, and every person to their role: managers see only the communities assigned to them, board members only their own community, and residents only their own home. Access checks run on every request, and row-level security on every database table blocks direct API access.
Humans stay in control
Montrux proposes; people approve. Outbound communication requires approval by default. Enforcement actions, fines and legal matters are never automated. High-risk messages — water, fire, injury, legal threats, fair-housing issues — are escalated to a manager immediately.
Grounded, verifiable answers
Answers about rules and responsibilities cite the community’s own documents. We check that every quoted passage actually exists in the source. When documents conflict or state law may apply, Montrux says so and defers to your team.
Untrusted input handling
Resident emails and uploaded documents are treated as data, never instructions. The AI cannot send messages to arbitrary recipients or take actions on its own; every action is validated against a fixed schema before it runs.
Complete audit trail
An append-only log records what came in, what the AI concluded, which sources it used, what it recommended, who approved, and what was sent. Sign-ins and any Montrux support access are logged too, and the database itself refuses to change or delete log entries.
Data handling
Data is encrypted in transit (TLS) and at rest. Files are never publicly addressable. We use Anthropic’s API for language models; API data is not used to train models. We don’t sell data. Customers can request export or deletion at any time.
Questions or a security review? Email security@montrux.ai.