Live site: https://mannyyang.github.io/clypfeed/
AI news digest CLI that ingests Gmail newsletters and RSS feeds, extracts AI/ML news, and produces daily digests using the Claude Agent SDK.
curl -sSL https://raw.githubusercontent.com/mannyyang/clypfeed/main/install.sh | bashOr manually:
git clone https://github.com/mannyyang/clypfeed.git
cd clypfeed
npm install- Node.js 22+
- Claude Code (uses your subscription — no API key needed)
To pull newsletters from Gmail, create a .env file:
EMAIL_USER=you@gmail.com
EMAIL_PASSWORD=your-app-password
To get an app password:
- Enable 2-Step Verification
- Go to App passwords and generate one
- Enable IMAP: Gmail → Settings → Forwarding and POP/IMAP → Enable IMAP
clypfeed <command> [args] [flags]
Commands:
run Generate a new digest now
list [--limit N] List available digests (default: 20)
get <date> Get digest for a specific date (YYYY-MM-DD)
search <query> [--days N] Search digests by keyword (default: 7 days)
pages Build HTML pages from digests
feed list List configured RSS feeds
feed add <name> <url> Add an RSS feed
feed remove <name> Remove an RSS feed by name
clypfeed run # generate today's digest
clypfeed list # see available digests
clypfeed get 2026-03-10 # read a specific digest
clypfeed search "Claude" --days 14 # search recent digests
clypfeed feed add "TechCrunch" "https://techcrunch.com/feed"Output is saved to output/YYYY-MM-DD.json. HTML pages are built to docs/.
- Fetches RSS feeds from 15 curated AI/tech sources
- Connects to Gmail via IMAP and fetches recent newsletters (if configured)
- Scrapes official sources (Anthropic newsroom, Claude Code releases)
- Claude Agent autonomously reads, prioritizes, and deduplicates stories
- Saves a structured JSON digest and builds HTML pages
CLI (src/cli.ts)
└── Claude Agent SDK (query)
└── In-process MCP Server ("gmail-tools")
├── fetch_emails — list recent emails via IMAP
├── get_email — read & parse a single email
├── fetch_rss — pull items from RSS feeds
├── fetch_webpage — scrape web sources
├── get_published — check previously published stories
└── save_digest — write JSON digest to disk
ISC