interester

A personalized content monitoring system that allows users to register interests (topics, terms, URLs) and receive AI-curated, formatted summaries from across the web.

Project Overview

Purpose: A personalized content monitoring system that allows users to register interests (topics, terms, URLs) and receive AI-curated, formatted summaries from across the web.

Core Value Proposition:

• Intelligent Search: Goes beyond simple keyword matching to find relevant content across the web.
• AI Curation: Uses LLMs (Google Gemini via Vercel AI SDK) to summarize and format content into digestible insights.
• Flexible Monitoring: Track specific terms, broad topics, or specific URLs.
• Privacy-Focused: Local-first architecture keeps your data on your machine.

---

Current Status: MVP (Phase 1+)

The project has moved beyond the initial MVP, now featuring robust background monitoring and result management.

Features Implemented

• Interest Management: Create, update, and delete interests with search terms and specific URLs to monitor.
• Dashboard: Visual overview of active interests and system status.
• Automated Monitoring: In-app scheduler triggers scans on hourly, daily, or weekly intervals (active while app is open).
• System Notifications: Native desktop notifications when new results are found according to your interests.
• Result Management: Detailed AI summaries with key points and sources, featuring read/unread status, archiving, and deletion.
• Result Notes: Attach personal thoughts or findings to specific AI-curated results.
• Manual Trigger: Ability to manually run any interest search at any time.
• Local Storage: All data (interests, results, notes) is stored locally in JSON files (data/ directory).
• AI Integration: Integration with Vercel AI SDK using Google Gemini for intelligent content processing.
• Desktop App: Native desktop application performance and integration via Tauri v2.

Technology Stack

• Frontend: SvelteKit (Svelte 5) + Vanilla CSS
• Backend Logic: Node.js (running within the Tauri context)
• Desktop Framework: Tauri v2
• AI/LLM: Vercel AI SDK (ai, @ai-sdk/google) using Google Gemini
• Search: Serper API
• Storage: Local JSON files via Tauri file system APIs

---

Getting Started

Prerequisites

• Node.js (v18+)
• Rust (for Tauri)
• Google Gemini API Key
• Serper API Key

Installation

1. Clone the repository

   git clone <repository-url>
   cd interester/interester

2. Install dependencies

   npm install

3. Set up environment variables Create a .env file in the interester directory:

   GEMINI_KEY=your_gemini_api_key_here
   SERPER_KEY=your_serper_api_key_here

4. Run the development server

   npm run tauri dev

---

UX Overhaul Plan

To transition from a "functional MVP" to a "power-user tool," the following UX improvements are planned:

• Dashboard Intelligence:
  • Add an "Activity Stream" showing recent hits across all interests.
  • Integration of "System Health" (last scan time, API status).
  • Quick-add input for new interests directly from the home screen.
• Visual Feedback & Polish:
  • Implementation of Skeleton Loaders for data-heavy views.
  • Pulsing status indicators for active background tasks.
  • Enhanced card states (hover interactions, transition animations).
• Result Interactivity:
  • "Pinning" important results to the top of an interest feed.
  • Inline quick actions (Mark as read, Archive) without entering the detail view.
  • Source rich previews (favicons and site meta-data).
• Onboarding Experience:
  • Pre-populated "Example Interests" for first-time users.
  • Guided empty states with helpful tips for getting started.
• Ignore Rules (Noise Control):
  • Dedicated UI for defining global and per-interest ignore patterns (domains, keywords).
• Dark Mode Support:
  • Native dark theme to match OS preferences.

---

Roadmap

Phase 1: MVP Polish (Completed)

• Basic CRUD for Interests
• Local JSON Storage
• Tauri App Shell
• Manual "Run Search" trigger
• Scheduling & Notifications
• Result Management & Notes

Phase 2: UX & System Polish (Upcoming)

• Implement UX Overhaul Plan (see section above)
• True Background Support: Migrate scheduling to Rust/Tauri to enable scans even when the app is closed.
• Refine AI summarization prompts for better consistency
• Improve search result parsing for broader site compatibility
• Build Ignore Rules UI for noise reduction

Phase 3: Web & Cloud (Future)

• Web Version: Extract backend to run as a standalone web server.
• Authentication: User accounts for the web version.
• Database: Migrate from JSON to a proper database (SQLite/Postgres).
• Multi-channel Delivery: Email, SMS, and Webhook notifications.