Open-source automated intelligence for personal and public safety.
Estate Sentry is a prototype threat intelligence platform that amplifies personal and public safety through automated sensor monitoring, zone-based perception, and AI-powered threat analysis. It is designed to be transparent, extensible, and self-hosted.
This is a demonstrational prototype under active development. It is not licensed for public intelligence gathering.
Cameras and sensors feed into a layered perception pipeline that detects objects, identifies people, recognizes actions, and correlates events across zones. Threat scores are computed in real time and high-severity events are analyzed by an LLM for natural-language reasoning.
Cameras & Sensors --> Switchboard (NATS) --> Perception Pipeline --> Threat Assessment
| |
Zone Events AI Analysis
Identity Store Alert System
| |
+-----> HQ Dashboard <---+
| Layer | Technology |
|---|---|
| API | Django 5 + Django REST Framework |
| Dashboard | Django templates + htmx (no build step) |
| Message Bus | NATS 2.10 (JetStream) |
| Relational DB | PostgreSQL (TimescaleDB) |
| Graph DB | Neo4j 5 |
| Vector DB | ChromaDB |
| Object Storage | MinIO |
| AI Analysis | Claude via MCP |
| Tooling | uv (Python), Task runner |
git clone https://github.com/QR-Madness/estate-sentry.git
cd estate-sentry
task setup
task api:superuser
task devAPI and dashboard at localhost:8000
For Docker deployment: task docker:up (starts PostgreSQL, Neo4j, NATS, MinIO, ChromaDB and the API).
Full documentation is published at QR-Madness.github.io/estate-sentry or serve locally with task docs:serve.
Key docs:
- Specification — Perception layers, zone model, identity pipeline, threat scoring
- Architecture — System design and database roles
- Roadmap — Implementation plan and milestones
See the Development Guide for tooling, testing, and contribution workflow.
See LICENSE for details.