Newsletter & Blog Digester

A small application that periodically checks websites/blogs, summarizes new posts using OpenAI, and sends digests to Slack.

Features & Benefits

• Multiple Extraction Methods: RSS feeds, CSS selector rules, or AI-powered extraction for complex HTML layouts
• Flexible AI Options: Use OpenAI API or run Ollama locally for completely offline operation
• Smart Content Processing: Automatic deduplication, summarization, and notification management
• Slack Integration: Multi-channel notifications with customizable bot name and emoji
• Single Docker Container: Everything runs in one lightweight container with auto-reload during development
• Zero Build Step Frontend: Preact + HTM loaded from CDN - edit and refresh, no compilation needed
• Fully Offline Capable: Works without internet when using local Ollama for AI processing
• Lightweight Stack: Node.js + Fastify + SQLite - minimal dependencies, fast startup, low resource usage
• Scheduled Monitoring: Configurable cron schedules with manual “Check Now” option
• Web UI: Clean interface for managing sites, viewing posts, configuring settings, and monitoring logs

Screenshots

Here is what it looks like:

Tech Stack

• Runtime: Node.js 20
• Framework: Fastify
• Database: SQLite (better-sqlite3)
• Frontend: Preact + HTM + snarkdown
• Testing: uvu + sinon + c8
• Formatting: dprint (Rust-based formatter) / 10-100x faster than Prettier / Zero npm dependencies
• Git Hooks: Husky
• Container: Docker Compose

Quick Start

Local Development

# Start
docker compose up -d

Server will be available at: http://localhost:5566

Testing

Tests use:

• uvu: Fast, lightweight test runner
• sinon: Mocking library for external APIs
• c8: Code coverage tool

Test files: src/server/__tests__/*.test.js

License

MIT