v4.4: webhook mode for Cloud Run / Railway / Render
ilang-ai
3e53fa5
- Auto-detect: WEBHOOK_URL set → webhook mode, otherwise → polling mode
- Webhook: listens on PORT (default 8080), registers /webhook endpoint
- Polling: unchanged behavior for VPS / local dev
- Dockerfile: PORT=8080 for Cloud Run default
- Health check only starts in polling mode (webhook has its own HTTP)
fix: HF Space health check on port 7860 + Node.js 24 Actions warning
ilang-ai
59295ba
- Add HTTP health check thread (port 7860) so HF Space doesn't restart container
- Add EXPOSE 7860 to Dockerfile
- Add FORCE_JAVASCRIPT_ACTIONS_TO_NODE24 to both workflows
v1.0: TelegramGuard - AI-powered group guardian with I-Lang Prompt Spec
ilang-ai
cc6f785
- Anti-spam: AI + repeat message detection (3x in 5min = ban)
- Vision: image and video thumbnail understanding
- Chat: group @mention and private conversation
- Prompts: .ilang format (I-Lang Prompt Spec)
- Deploy: HF Space (free) or VPS
- GitHub Actions: auto-sync to HuggingFace
