Instapaper Management App

A web application for managing your Instapaper bookmarks with a clean interface and keyboard shortcuts. The nice thing is this stores past the 500 article limit you normally get back from the API

Features

• View all Instapaper bookmarks in a list
• Read bookmark content directly in the app
• Archive bookmarks
• Star/unstar bookmarks
• Track reading progress
• Keyboard navigation
• SQLite storage for offline access
• AI-powered semantic search using embeddings
• Automatic article summarization using GPT-3.5
• RAG (Retrieval Augmented Generation) for efficient content processing

Setup

1. Install dependencies:

npm install

2. Create a .env file with your Instapaper credentials:

INSTAPAPER_CLIENT_ID=your_client_id
INSTAPAPER_CLIENT_SECRET=your_client_secret
INSTAPAPER_EMAIL=your_email
INSTAPAPER_PASSWORD=your_password
OPENAI_API_KEY=your_openai_api_key

3. Create SQLite database schema:

CREATE TABLE articles (
    title TEXT,
    url TEXT,
    text TEXT,
    hash TEXT,
    bookmark_id INTEGER PRIMARY KEY,
    time INTEGER,
    progress_timestamp INTEGER,
    progress REAL,
    archived BOOLEAN,
    summary TEXT
);

CREATE TABLE article_chunks (
    bookmark_id INTEGER,
    chunk_index INTEGER,
    chunk_text TEXT,
    embedding BLOB,
    PRIMARY KEY (bookmark_id, chunk_index)
);

4. Populate DB

node load-articles.js

Usage

Start the server:

node index.js

The app will be available at http://localhost:3000

Keyboard Shortcuts

• j - Move selection down
• k - Move selection up
• l - Toggle star
• Enter - View article text
• Backspace - Archive article
• 0 - Jump to top
• r - Select random article
• s - Generate/show article summary

Architecture

• Backend: Express.js server with SQLite database
• Frontend: Svelte application
• API: Instapaper SDK for Node.js
• AI: OpenAI API integration for embeddings and summarization
• RAG: Custom service for content processing and semantic search

Data Flow

1. Initial load syncs bookmarks from Instapaper to local SQLite database
2. Articles are processed through RAG service for embeddings and summaries
3. Frontend communicates with Express backend for data operations
4. Changes are persisted to both local database and Instapaper through the API
5. Search queries are processed using semantic embeddings for better results