Building AI-powered tools and platforms? Let's talk about your next project.
hello@harborsoftware.com
Case Study
ShelfMark
AI Bookmark Auto-Categorizer
Overview
ShelfMark is a Chrome Extension that auto-categorizes bookmarks into native browser folders. It uses a hybrid approach: fast rule-based domain matching for obvious categorizations, with LLM fallback via OpenRouter for ambiguous content.
The Challenge
Browser bookmarks become an unmanageable mess within weeks. Manual folder organization doesn't scale, and existing bookmark managers require switching to a separate app. Users need categorization that works inside the browser, automatically.
Key Features
- Automatic categorization into native Chrome bookmark folders
- Rule-based domain matching (fast, deterministic)
- LLM fallback for ambiguous bookmarks via OpenRouter
- Privacy-first — data stays in browser unless LLM is explicitly enabled
- Bulk import and retroactive organization
- Folder conflict resolution with existing user structure
Tech Stack
- React 18
- TypeScript
- Vite
- Tailwind CSS v4
- shadcn/ui
- Chrome Extension APIs
- OpenRouter
Status
Production Chrome Extension