Amnesia

Secure visual asset scrubbing and privacy remediation protocol.

Amnesia identifies and remediates visual identifiers that link digital assets to physical locations. It uses high-inference scrubbing to wipe regional traces, landmarks, and boutique signatures that can be used for geolocation.

Features

Feature	Description
Visual Scanning	Scans URLs for images and videos containing potential privacy leaks.
High-Inference Analysis	Detects landmarks, street signs, reflections, and regional foliage via Gemini.
Zero-Cloud Protocol	Toggle local VLM mode to run scans entirely offline (requires Ollama/LLaVA).
Sanitizer UI	Interactive "Wipe & Download" redaction to apply heavy visual blurs before saving.
Ghost Mode Extension	Right-click any web image to send it directly to your local Amnesia instance.
CLI Scrubber	Batch metadata erasure for local visual assets.

Installation

Build from Source

git clone https://github.com/ghostintheprompt/amnesia.git
cd amnesia
npm install

Usage

1. Web Interface:

   • Run npm run dev to start the local server.
   • Enter a target URL to scan for assets.
   • Run "Wipe Analysis" on individual assets to detect privacy risks, or use the "Zero-Cloud" local mode.
   • Click "Wipe & Download" to save a visually sanitized version of the image.

2. Ghost Mode Browser Extension:

   • Go to chrome://extensions/ and load the unpacked extension/ folder.
   • Right-click any image or page and select "Send to Amnesia Scrubber".

3. Global CLI Scrubber:

   • Install globally: npm install -g .
   • Run amnesia-scrub <directory_path> anywhere to strip all metadata from local assets.

Privacy Statement

Amnesia is built by MDRN Corp. It operates 100% locally or via direct API calls to Google Gemini. No telemetry, no tracking, no bloat.

---

Built by MDRN Corp — mdrn.app