Project Overview
LeakShot is a local-first screenshot privacy scanner that helps users find and redact sensitive text before sharing an image. It keeps the workflow private by processing screenshots in the browser without uploads or accounts.
Problem
Screenshots often contain emails, addresses, API keys, tokens, and private notes. Users usually notice these details too late or rely on manual markup tools that are slow and easy to misuse.
Solution
LeakShot scans an uploaded screenshot locally, highlights risky text regions, and gives users a redaction preview before exporting a clean version.
Key Features
- 100% local screenshot processing
- Sensitive text detection
- Redaction preview before export
- Clean image export
Technical Decisions
The product uses a Next.js and TypeScript frontend with OCR-driven text detection and client-side image handling. The local-first structure avoids a backend for the core privacy workflow and keeps the app fast to launch.
UI/UX Decisions
The interface is designed around a short scan, review, export flow so users can understand risk quickly and leave with a clean image in a few steps.
Challenges
The hardest part is balancing detection quality with speed while keeping the workflow understandable for non-technical users.
Result
Users can inspect screenshots, catch sensitive details, redact risky areas, and export a safer image without sending private content to a server.
Screenshots
