Never miss a self-hosted software update again
Complete workflow JSON with 9 sticky notes — import into n8n and connect credentials
PostgreSQL table for release history with indexes for analytics
8-panel analytics dashboard for release tracking and security monitoring
Minimal n8n + PostgreSQL setup for testing the workflow standalone
Immich, Jellyfin, Vaultwarden, Traefik, Uptime Kuma, n8n, Watchtower, Nextcloud, Grafana, Prometheus
Built-in test mode fires a sample alert through the full pipeline — verify everything works before going live
9-step README covering import, credentials, migration, configuration, and activation
.env.example with all configurable variables documented
GitHub Releases API, Docker Hub tags, and GHCR tags — all in one workflow.
Claude Haiku summarizes changes, flags breaking changes, detects CVEs, and rates urgency 1-5.
Discord webhooks, Telegram bot, Slack app, and ntfy push notifications. Enable any combination per repo.
Point to your docker-compose.yml and every service gets watched automatically. Supports 25+ common images.
Critical and security items sent immediately. Routine updates batched into a daily digest.
Custom channels, urgency overrides, and instant vs batch delivery for each monitored repo.
Enabled by default — click Test to verify the full pipeline (AI analysis, formatting, delivery) without waiting for a real release.
Claude Haiku processes changelogs cheaply. Most days cost under a penny.
An n8n workflow that monitors GitHub repos and container registries (Docker Hub, GHCR) for new releases. Uses Claude AI to summarize changelogs, flag breaking changes, detect CVEs, and rate urgency. Delivers color-coded digests to Discord, Telegram, Slack, and ntfy with smart routing.
Built for homelabbers who run self-hosted software and want to stay on top of updates without manually checking each repo.
## How It Works
```
Schedule (8 AM daily) or Manual Trigger
→ Configure Watcher (all settings in one node)
→ Build Repo Watchlist (manual repos + auto-detected from docker-compose)
→ Source Router: GitHub → Releases API, Registry → Tags API
→ Merge all sources → Compare Versions → Filter new only
→ Claude AI: analyze changelog, flag security, rate urgency
→ Save to PostgreSQL (parallel)
→ Apply per-repo alert rules
→ Urgency Router: critical/instant → immediate alert, batch → daily digest
→ Channel Router → Discord, Telegram, Slack, ntfy (per repo config)
→ Update stored version tags
```
## Setup (~10 minutes)
1. Import `workflow.json` into n8n
2. Add credentials (Anthropic API key, PostgreSQL, notification channels)
3. Run the database migration (`migrations/001_release_history.sql`)
4. Configure repos and channels in the Configure Watcher node
5. Test with built-in test mode (fires a sample alert through the full pipeline)
6. Switch to production and activate
## Pre-loaded Repos
Comes with 10 popular homelab repos pre-configured: Immich, Jellyfin, Vaultwarden, Traefik, Uptime Kuma, n8n, Watchtower, Nextcloud, Grafana, and Prometheus.
## Docker Compose Auto-Detection
Point to your `docker-compose.yml` and every service gets watched automatically. Supports 25+ common homelab images. Manual repos always take priority over auto-detected duplicates.
## Alert Rules
Each repo can configure which channels receive alerts, whether to send immediately or batch into a daily digest, and optionally override the AI urgency rating.
## GitHub Token
Without a token: 60 requests/hour (fine for 10-15 repos). With a token: 5,000 requests/hour (recommended for 15+ repos).
## Grafana Dashboard
Includes an 8-panel Grafana dashboard for release analytics: Releases Over Time, Breaking Changes, Security Advisories, Releases by Source, Most Active Repos, Urgency Distribution, Recent Releases table, and Total Releases Tracked.
- n8n v1.0+ (self-hosted)
- Anthropic API key (Claude Haiku, ~$0.01-0.03 per run)
- PostgreSQL database (for release history)
- At least one delivery channel: Discord webhook, Telegram bot, Slack app, or ntfy topic
- Optional: GitHub Personal Access Token (for higher rate limits with 15+ repos)
- Optional: Grafana (for the analytics dashboard)
Learn how this was built, the decisions made, and gotchas to watch for.
Patterns, lessons, and tools from production automation systems. Delivered weekly.