Repository navigation
Releases: Farukk57/release-notifier
Release list
v1.3.0 — OpenAI Support, HA Notifications & Security Hardening
What's new in v1.3.0
OpenAI support — use GPT-4o-mini or GPT-4o as the AI provider.
AI_PROVIDER=openai
OPENAI_API_KEY=sk-...
OPENAI_MODEL=gpt-4o-miniHome Assistant persistent notifications — release summaries now also appear in the HA bell icon as browsable history, not just as mobile push notifications.
Bug fixes & security hardening:
- Fixed HA mobile push notifications missing section headers (What's new / Fixes / Action required)
- Atomic file writes — prevents state.json/notified.json corruption on container restart
- GitHub repo field validation in containers.yml
- Home Assistant notify service name validation
- Pre-compiled regex patterns for better performance
- Updated Claude model to
claude-sonnet-4-6
Full changelog: v1.2.0...v1.3.0
v1.2.0 — Home Assistant Integration
What's new in v1.2.0
Home Assistant integration — send rich push notifications directly to your phone via the HA companion app alongside or instead of ntfy.
- Mobile push notification with urgency-based priority
- Persistent notification history in the HA bell icon — stays until dismissed
- Critical alerts — urgent releases (CVEs) bypass Do Not Disturb on iOS
- Tap the notification to open the GitHub release page directly
Setup: Set NOTIFY_PROVIDER=both, HASS_URL, HASS_TOKEN, and HASS_NOTIFY_SERVICE=mobile_app_your_phone in your .env.
Full changelog: v1.1.0...v1.2.0
v1.1.0 — Ollama Support & Permission Fix
What's new in v1.1.0
Ollama support — run AI summaries locally for free using any Ollama model. No Anthropic API key needed. Set AI_PROVIDER=ollama, OLLAMA_URL=http://your-host:11434 and OLLAMA_MODEL=mistral in your .env.
Recommended models: mistral (best quality), llama3.2 (faster, smaller)
Note: Ollama responses take 10–60 seconds per summary depending on your hardware vs ~3 seconds with Claude. Urgency classification may be less accurate with smaller models.
Volume permission fix — container now runs as UID 1000, so mounted config volumes work out of the box on standard Linux without needing chmod.
Full changelog: v1.0.0...v1.1.0
v1.0.0 — Initial Release
🚀 Initial Release
release-notifier watches your self-hosted Docker stack for new GitHub releases, summarises the changelog with Claude AI, and pushes a notification to your phone via ntfy.
Features
- 🔔 Push notifications via ntfy with urgency-based priority (urgent for CVEs, high for breaking changes)
- 🤖 AI-powered release summaries using Claude -> what changed, security patches, action required
- 🐳 Registry digest tracking -> knows whether you've already pulled the new image or not
- 🏷️ OCI label version detection -> shows
v1.0.0 -> v1.1.0in notification titles - 🔍 Pre-release filtering -> ignores nightlies and RCs by default
- 🔁 Deduplication -> won't re-notify for versions already sent
- 🔄 Retry with exponential backoff on transient failures
- 📊 Homepage iframe dashboard widget
- 🔒 Production-ready: waitress WSGI, non-root container, XSS protection, optional API key auth
Quick start
docker compose -f docker-compose.hub.yml up -dSee the README for full setup instructions.