I built an OSS newsletter digester that uses AI to send me daily Slack summaries (github.com)

A developer released an open-source "newsletter digester" that periodically crawls sites and blogs, summarizes new posts using AI, and delivers daily Slack digests. It supports multiple content extraction methods — RSS, CSS-selector rules, or an AI-powered scraper for messy HTML — and performs automatic deduplication, summarization, and notification management. You can use the OpenAI API or run Ollama locally for fully offline summarization, making it useful for privacy-conscious teams or environments without internet access. The app is deliberately lightweight and developer-friendly: it runs in a single Docker container (Node.js 20 + Fastify) with SQLite (better-sqlite3) persistence, a zero-build Preact + HTM frontend loaded from CDN, cron-configurable checks, and manual "Check Now" triggers. Slack integration supports multi-channel notifications with customizable bot identity. The repo includes testing (uvu, sinon, c8), fast formatting with dprint, Husky git hooks, and auto-reload during development. MIT-licensed and easy to spin up (docker compose up -d → http://localhost:5566), it’s practical for AI/ML teams to automate content curation, reduce noise, and keep distributed teams synced on relevant OSS/ML updates without heavy infrastructure.