# Sonomos — Privacy Layer for AI

> Thoughts belong to thinkers.

## What is Sonomos?

Sonomos is a browser extension that protects sensitive information when using AI assistants like ChatGPT, Claude, Gemini, and others. It runs entirely in your browser—no servers, no network calls, no data leaving your machine.

## How it works

1. **Dagger (Detection)**: Real-time detection of sensitive data as you type. Underlines flagged content with color-coded risk levels (yellow = medium, red = high).

2. **Cloak (Protection)**: Transforms sensitive data before it reaches AI systems using obfuscation and tokenization while preserving meaning.

3. **Traffic Light Widget**: Visual indicator showing current protection status (green = safe, yellow = review recommended, red = high risk).

## Key Facts

- **Architecture**: 100% client-side, local-first browser extension
- **Privacy model**: Zero-knowledge, on-device processing
- **No data collection**: Nothing is stored, logged, or transmitted
- **Supported platforms**: Works with ChatGPT, Claude, Gemini, Grok, and other AI interfaces
- **Target users**: Privacy-sensitive professionals (lawyers, consultants, regulated industries)

## What Sonomos is NOT

- Not a chatbot or AI wrapper
- Not a cloud service that processes your data
- Not a marketing tool or analytics platform
- Does not require an account to function

## Core Pages

- Homepage: https://sonomos.ai/
- Product details: https://sonomos.ai/product
- Mission statement: https://sonomos.ai/mission
- Team: https://sonomos.ai/team
- Blog: https://sonomos.ai/blog
- FAQ: https://sonomos.ai/faq
- Documentation: https://sonomos.ai/docs
- Contact: https://sonomos.ai/contact

## Legal & Security

- Privacy Notice: https://sonomos.ai/privacy
- Security: https://sonomos.ai/security
- Terms of Service: https://sonomos.ai/terms
- No-Tracking Policy: https://sonomos.ai/no-tracking

## Company

Sonomos, Inc.
Website: https://sonomos.ai
Contact: hello@sonomos.ai

## Technical Summary

Sonomos intercepts text input in AI chat interfaces before submission. Sensitive data patterns (names, addresses, financial info, medical terms, credentials) are detected using local pattern matching. Users can choose to mask, replace, or remove flagged content. The AI receives only the sanitized version. Original content never leaves the browser.

---

Last updated: 2026-02-04