Security & Responsible AI
Trust is built
piece by piece.
Our intended approach to data and AI, with a clear line between principles and policies still being finalized.
01
Collect deliberately.
Our goal is to request only the data a task needs and explain why it is needed. Product-specific collection, storage locations, access controls, deletion options, and retention periods are not yet finalized.
This website has no analytics or advertising trackers. Forms only deliver information when a real submission service is connected. See the draft Privacy Policy for the proposed broader framework.
02
Training choices should be clear.
Our intended policy is not to use customer content to train models by default. This is pending confirmation, not a live-product guarantee.
Before product access, we need to confirm provider settings, contractual restrictions, and any optional data-sharing controls. Finalized terms will explain the choices available.
03
Know who processes the data.
AI model providers and other product subprocessors have not been finalized. We will need to disclose relevant providers, their roles, and material processing terms before asking customers to use a product.
We do not claim zero retention, a particular hosting region, or a specific provider's protections at this stage.
04
Keep people in control.
We're designing toward visible drafts and reviewable actions. People should be able to question a result and decide what happens next.
AI can be inaccurate, incomplete, or biased. Outputs need appropriate review; our products are not intended to replace professional medical, legal, or financial judgment.
05
Make room for responsible reporting.
If you believe you've found a security issue, email a concise description and safe reproduction steps. Please do not send private customer data or exploit a suspected issue beyond what is necessary to describe it.
We do not currently publish a bug bounty program, guaranteed response time, or compliance certification.
Report a security concern
A direct line for security.
Share enough detail to help us understand the issue, without exposing anyone else's information.