Privacy-First, Self-Hosted, Kubernetes-Deployed AI Assistant
Kilo Guardian is a comprehensive cognitive support system running entirely on your local infrastructure. It combines AI-powered memory management, health tracking, financial oversight, and habit formation into a unified, privacy-first platform.
Current Status: β 100% Operational - 15 microservices running on K3s
# SSH tunnel to access Kilo Guardian
ssh -L 3000:localhost:30000 -L 8000:localhost:30800 kilo@192.168.68.66Then open: http://localhost:3000
See docs/TABLET_ACCESS.md for detailed setup.
- Frontend: http://localhost:30000
- Gateway API: http://localhost:30800
βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ
β KILO GUARDIAN KUBERNETES CLUSTER β
β (K3s on Pop!_OS) β
βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ
External Access (NodePort):
βββΊ Frontend (30000) βββΊ React UI
βββΊ Gateway (30800) βββΊ API Router
Kubernetes Services (ClusterIP):
βββΊ Frontend Service : kilo-frontend (80)
βββΊ API Gateway : kilo-gateway (8000)
β
βββΊ Core Services:
β βββΊ Medications : kilo-meds (9001)
β βββΊ Medications v2 : kilo-meds-v2 (9001)
β βββΊ Reminders : kilo-reminder (9002)
β βββΊ Habits : kilo-habits (9003)
β βββΊ Financial : kilo-financial (9005)
β βββΊ Library : kilo-library (9006)
β
βββΊ Intelligence Layer:
β βββΊ AI Brain : kilo-ai-brain (9004)
β βββΊ ML Engine : kilo-ml-engine (9008)
β βββΊ Ollama : kilo-ollama (11434)
β
βββΊ I/O Services:
β βββΊ Camera : kilo-cam (9007)
β βββΊ Voice : kilo-voice (9009)
β βββΊ USB Transfer : kilo-usb-transfer (8006)
β
βββΊ Real-Time:
βββΊ SocketIO : kilo-socketio (9010)
All services in namespace: kilo-guardian
Network: 10.42.0.0/16 (K3s Pod Network)
Infrastructure:
- K3s (Lightweight Kubernetes)
- Kubernetes 1.28+
- Pop!_OS 22.04 LTS
Backend:
- Python 3.11
- FastAPI
- SQLite + SQLModel
- sentence-transformers
- Ollama (Local LLM)
Frontend:
- React 19.2.3
- TypeScript 4.9.5
- TailwindCSS
- React Router v6
| Service | Pod Name | Status | Function |
|---|---|---|---|
| Frontend | kilo-frontend | β Running | React UI |
| Gateway | kilo-gateway | β Running | API Router & Auth |
| Medications | kilo-meds | β Running | Med tracking & OCR |
| Medications v2 | kilo-meds-v2 | β Running | Updated version |
| Reminders | kilo-reminder | β Running | Timeline & alerts |
| Habits | kilo-habits | β Running | Habit tracking |
| AI Brain | kilo-ai-brain | β Running | RAG & Memory |
| Financial | kilo-financial | β Running | Budget & receipts |
| Library | kilo-library | β Running | Knowledge base |
| Camera | kilo-cam | β Running | Pose detection |
| ML Engine | kilo-ml-engine | β Running | ML processing |
| Voice | kilo-voice | β Running | Voice input |
| USB Transfer | kilo-usb-transfer | β Running | File transfer |
| SocketIO | kilo-socketio | β Running | Real-time events |
| Ollama | kilo-ollama | β Running | Local LLM |
Total: 15 pods, all healthy
- β 100% Self-Hosted - All data stays on your server
- β No Cloud Dependencies - Fully offline capable
- β Local AI - Ollama runs LLMs on-premise
- β Encrypted Storage - Sensitive data protected
- β Network Isolation - K3s internal networking
- β Semantic Memory - RAG-powered context recall
- β Smart Suggestions - AI-driven recommendations
- β Natural Language - Chat interface for all modules
- β Context Awareness - Learns your patterns
- β Touch-Friendly UI - Large touch targets (60px+)
- β Responsive Design - Works on any screen size
- β PWA-Ready - Install as app on mobile
- β Fast Performance - Optimized React build
- β High Availability - K3s self-healing
- β Service Discovery - Automatic DNS routing
- β Health Monitoring - Built-in health checks
- β Easy Scaling - Kubernetes-native scaling
- β Rolling Updates - Zero-downtime deployments
- TABLET_ACCESS.md - Access from tablet/mobile
- K3S_ACCESS_GUIDE.md - Kubernetes deployment guide
- DEPLOYMENT_GUIDE.md - Full deployment instructions
- POD_HEALTH_REPORT.md - Current system status
- SERVICE_COMMUNICATION_TEST.md - Connectivity verification
- K8S_HARDENING_SUMMARY.md - Security configuration
- ROADMAPS/INTEGRATION_ROADMAP.md - Future integration plans
- ROADMAPS/VOICE_ROADMAP.md - Voice feature roadmap
- EXTERNAL_CAMERA_IMPLEMENTATION.md - Camera system
- MULTI_CAMERA_SYSTEM.md - Multi-camera setup
- PERFORMANCE_IMPROVEMENTS.md - Optimization history
# View all pods
kubectl get pods -n kilo-guardian
# Check services
kubectl get svc -n kilo-guardian
# View logs for a service
kubectl logs -n kilo-guardian deployment/kilo-gateway --tail=50# Restart a service
kubectl rollout restart deployment/kilo-meds -n kilo-guardian
# Scale a service
kubectl scale deployment/kilo-ml-engine --replicas=2 -n kilo-guardian
# View resource usage
kubectl top pods -n kilo-guardian# Check pod details
kubectl describe pod <pod-name> -n kilo-guardian
# Get pod events
kubectl get events -n kilo-guardian --sort-by='.lastTimestamp'
# Access pod shell
kubectl exec -it deployment/kilo-gateway -n kilo-guardian -- /bin/shSee docs/OPERATIONS.md for comprehensive operations guide.
kilo-guardian-core/
βββ services/ # Microservice implementations
β βββ ai_brain/ # RAG & memory search
β βββ cam/ # Camera & pose detection
β βββ financial/ # Budget & transaction tracking
β βββ gateway/ # API router & authentication
β βββ habits/ # Habit tracking & analytics
β βββ library_of_truth/ # Knowledge base & PDF storage
β βββ meds/ # Medication management
β βββ ml_engine/ # ML processing engine
β βββ reminder/ # Timeline & reminders
β βββ socketio-relay/ # Real-time communication
β βββ usb_transfer/ # File transfer service
β βββ voice/ # Voice input processing
β βββ k3s_manager/ # K3s cluster management
β
βββ frontend/ # React frontend
β βββ kilo-react-frontend/
β βββ src/
β β βββ components/
β β βββ pages/
β β βββ services/
β βββ public/
β
βββ core/ # Core agent functionality
β βββ kilo_agent.py # Main agent logic
β βββ kilo_agent_brain.py
β βββ kilo_agent_ui.py # UI interface (canonical)
β βββ kilo_memory.py # Memory management
β βββ kilo_router.py # Command routing
β βββ kilo_pod_access.py
β
βββ shared/ # Shared utilities & config
β βββ config.py # Centralized configuration
β βββ db.py # Unified database access (SQLModel)
β βββ models/ # Shared data models
β βββ utils/ # Helper functions
β
βββ k3s/ # Kubernetes manifests
β βββ deployments/
β βββ services/
β βββ configmaps/
β
βββ docs/ # Documentation
β βββ ROADMAPS/ # Future planning
β βββ REPORTS/ # Historical reports
β βββ *.md # Current documentation
β
βββ scripts/ # Operational scripts
β βββ deploy-to-k3s.sh
β βββ check-status.sh
β βββ launch_*.sh
β
βββ legacy/ # Legacy/experimental code
β βββ v1/ # Old version 1
β βββ v2/ # Old version 2
β βββ *.py # Archived experiments
β
βββ kilo_agent_api.py # Agent API service
βββ kilo_proactive_agent.py # Proactive monitoring agent
βββ start-proactive-agent.sh # Agent startup script
- Medication schedule with timers
- Prescription OCR scanning
- Dosage tracking
- Prescriber management
- Timeline view
- Voice input support
- Recurring reminders
- Priority levels
- Budget tracking
- Receipt OCR
- Transaction categorization
- Monthly summaries
- Goal setting
- Daily habit tracking
- Streak counters
- Progress visualization
- Weekly calendar view
- Custom icons
- Semantic memory search
- RAG-powered chat
- Context-aware responses
- Memory consolidation
- PDF knowledge base
- Document search
- Note management
- Tag organization
- Pod Startup: < 30 seconds
- API Response: < 100ms (avg)
- Frontend Load: < 2 seconds
- Memory Usage: ~4GB total
- CPU Usage: < 20% (idle)
- β Network Policies - Service-to-service restrictions
- β RBAC - Role-based access control
- β Pod Security - Non-root containers
- β Secret Management - Kubernetes secrets
- β Internal DNS - ClusterIP-only backend services
- β NodePort Limited - Only frontend & gateway exposed
# Test gateway
curl http://localhost:30800/meds/
# Test financial summary
curl http://localhost:30800/financial/summary
# Test reminder list
curl http://localhost:30800/reminder/remindersOpen http://localhost:30000 and verify:
- β Dashboard loads
- β All 6 modules accessible
- β Data persists across refreshes
- β Navigation works smoothly
- Check pod status:
kubectl get pods -n kilo-guardian - View logs:
kubectl logs <pod-name> -n kilo-guardian - Check events:
kubectl get events -n kilo-guardian
- Verify NodePort:
kubectl get svc -n kilo-guardian - Check firewall:
sudo ufw status - Test locally:
curl http://localhost:30000
- Check pod restart count:
kubectl get pods -n kilo-guardian - View logs for errors
- Verify volume mounts:
kubectl describe pod <pod-name> -n kilo-guardian
Contributions welcome! Please read CONTRIBUTING.md for guidelines.
MIT License - see LICENSE for details.
- K3s - Lightweight Kubernetes
- Ollama - Local LLM runtime
- sentence-transformers - Semantic embeddings
- FastAPI - Modern Python API framework
- React - UI framework
- TailwindCSS - Utility-first CSS
For issues or questions:
- Check logs:
kubectl logs <service> -n kilo-guardian - View documentation:
docs/ - System status: docs/POD_HEALTH_REPORT.md
β Infrastructure: K3s cluster fully operational β Backend: 13 microservices running β Frontend: React UI deployed and accessible β Database: SQLite with persistent storage β AI: Ollama LLM ready β Networking: All services communicating β Documentation: Comprehensive guides available
System Health: 100% - All 15 pods running
Built for privacy-conscious users who want powerful AI without compromising data sovereignty
π€ Deployed with Kubernetes | π Secured by Design | π Runs Entirely On-Premise