Privacy principles
Context without surrendering control.
Katally works with deeply personal information. Trust cannot be a feature added at the end — it has to shape the hardware, software, data model, and business model from the beginning.
Visible by default
A device that hides its state cannot be trusted.
Every Katally prototype has carried a display, because the people around a wearer deserve to know as much as the wearer does. The indicator is not a settings screen buried three taps deep — it is the front of the device.
Our design commitments
Clear boundaries for deeply personal technology.
These principles describe the product direction. Detailed policies will evolve with the product, infrastructure, and regions where Katally becomes available.
Sensing should be visible.
A user and the people around them should be able to understand when a Katally device is sensing, paused, charging, or synchronizing.
Pause should be immediate.
Privacy cannot depend on finding the right menu. Direct controls should make it possible to stop sensing in the moment.
Raw media and useful context are different.
Katally is being designed to let users choose whether raw media is retained, deleted after processing, or reduced to selected text context.
Personal data should stay portable.
Long-term context becomes more valuable over time. The product direction prioritizes exportable knowledge and user ownership.
AI training requires explicit consent.
Personal context should not be treated as a default training resource. Any future model-improvement use must be clear and optional.
More context does not always mean better support.
The system should help users define boundaries by time, place, activity, person, and retention period.
Architecture direction
Keep raw data close. Keep lasting context intentional.
The planned architecture separates local media capture, synchronization, AI processing, structured context, and long-term memory. That separation creates room for different privacy modes instead of one all-or-nothing cloud pipeline.
Local device storage
Selected AI analysis
Retain, reduce, or delete
Structured personal context
A conversation, not a checkbox
Help us build a privacy model worthy of real life.
We want early users, researchers, families, and developers to challenge these assumptions before the product reaches broad availability.
Challenge our assumptionsWrite to hello@katally.com — we read every message.