DEFINITION
Zero-Knowledge Processing
An architecture where the service provider has no technical ability to read user data, because all sensitive processing happens on the client.
In depth
Zero-knowledge processing means the service operator cannot see your data, even if compelled to. The pattern relies on doing all sensitive work on the client side — your browser, your device — and either never transmitting raw data or transmitting only opaque ciphertext that the server cannot decrypt. Compared to standard cloud processing, the trust boundary shrinks to your own machine. There is no server to breach, subpoena, or train on.
Examples
- Browser-based password managers that derive keys locally and store only encrypted blobs.
- End-to-end encrypted messengers where the server routes ciphertext it cannot decrypt.
- Local AI privacy layers (like Locke) that detect and transform sensitive data before any prompt leaves the device.
How Locke handles zero-knowledge processing
Locke is zero-knowledge by construction. Detection patterns ship with Locke and run on your machine. Locke makes no network calls during detection. Sonomos has no server that sees your prompts, your matches, or your masked values — there is nothing to log, nothing to leak.
See the productRelated terms
- On-Device AI
Running model inference or pre-processing entirely on the user's machine instead of sending data to a cloud service.
- Privacy Layer for AI
A control point between a user and an AI service that detects, transforms, or blocks sensitive data before it reaches the model.
- Tokenization
Replacing a sensitive value with a non-sensitive token that maps back to the original through a separate, controlled lookup.
