Skip to main content

Security and trust

Security information you can verify.

This page documents the controls and operating boundaries currently reflected in the Contextter product and public policies.

We do not claim certifications, residency commitments, or uptime guarantees that have not been independently documented.

The path of a request

Every protected request passes the same checkpoints before workspace data becomes reachable.

  1. Request

    Every access to protected areas starts without identity and without data access.

  2. Identity

    Clerk verifies identity and session. Without a signed-in user, the request ends here.

  3. Organization boundary

    Organization context is derived server-side from the authenticated session — never from client-supplied values.

  4. Workspace data

    Only after organization and workspace checks does the server return data for that workspace.

Authentication

The authenticated application uses Clerk for identity and session management. Protected application access requires an authenticated user.

Organization access

Workspace data is scoped through organization and workspace access checks. Public database functions derive organization context from authenticated access rather than trusting a client-supplied organization ID.

Roles and permissions

Access inside an organization is governed by graded roles — from owner and admin roles to restricted editing and viewing roles. Server-side checks are authoritative; interface elements are not a security boundary.

Uploads and customer data

Uploaded documents and the content derived from them are scoped to their workspace and are not available to other accounts.

AI processing

AI features run through a controlled internal layer that binds every call to user, organization, and workspace context and makes scope and cost traceable. Feature code does not talk to models directly.

Data and services

Contextter uses contracted infrastructure and service providers to operate authentication, storage, databases, billing, analytics, and selected product data workflows. Categories and purposes are listed in the privacy policy.

View providers

Secrets and logging

Credentials and API keys stay server-side. The logging layer automatically redacts sensitive fields; the content of contact requests is not logged.

Security-relevant development

Changes to authentication, permissions, billing, and public interfaces go through dedicated review. Data-model changes are additive and backwards compatible; tests protect auth and billing boundaries in proportion to risk.

What we deliberately do not claim

This page names no certifications, penetration-test reports, uptime guarantees, or compliance seals. If that changes, we will document it here with evidence.

Report a security concern

Include the affected URL or product area, a concise reproduction, observed impact, and a safe way to contact you. Do not include secrets or customer data in the first message.

Send a report

Questions and requests

Security, privacy, deletion, and access questions can be sent through the public contact channel for review and follow-up.