Privacy-first. Local-first. Signal over noise.
HNIS is a personal OSINT and safety monitoring platform that watches public surfaces, processes incoming signals, filters noise, and delivers actionable intelligence — all without sending sensitive data to the cloud.
Built by someone who depends on technology working reliably and privately, for people who feel the same way.
Most monitoring and alerting tools require you to either trust a third-party cloud with your data, manually check sources yourself, or accept a flood of noise with no intelligent filtering.
HNIS solves all three. It runs locally, watches automatically, thinks before it alerts, and delivers only what actually matters — directly to your device.
- Monitors public web surfaces, safety feeds, and configurable OSINT sources continuously
- Processes incoming signals through a local AI layer that understands context, not just keywords
- Filters aggressively — most noise never reaches you
- Delivers clean, actionable alerts locally to your device on your schedule
- Visualizes activity through a local dashboard designed with accessibility in mind
HNIS is built around a deliberate hybrid AI routing strategy:
| Task | Model | Reason |
|---|---|---|
| Complex reasoning, summarization, contextual analysis | Claude API | Accuracy and depth |
| All sensitive/personal data processing | Local Ollama instance | Privacy sovereignty — data never leaves your machine |
This means you get the capability of frontier AI where it counts, without sacrificing privacy where it matters.
- Backend: Python · FastAPI
- AI Layer: Claude API · Ollama (local)
- Data: PostgreSQL
- Dashboard: Local visual interface, screen reader accessible
- Alerting: Local delivery to iOS (iPhone)
- Architecture: Local-first · Privacy-first · No required cloud dependency
Privacy sovereignty — Sensitive data is processed exclusively on-device via local Ollama. Nothing personal touches an external API.
Signal over noise — The system is designed to be quiet. An alert means something. Noise is filtered before it reaches you.
Accessibility by default — The dashboard and alert delivery are built with screen reader compatibility from the ground up, not retrofitted.
Self-contained operation — Once deployed, HNIS runs without ongoing manual intervention. It watches so you don’t have to.
- JWT authentication on all API endpoints
- CORS protection
- No hardcoded credentials — environment-based configuration via central
core/config.py - All sensitive processing air-gapped to local model layer
- XSS protections in dashboard layer
🔧 Active development — Core pipeline, dashboard, and alert delivery operational. Ongoing refinement of signal filtering and source coverage.
- Expanded public source coverage
- Configurable alert thresholds per source type
- Enhanced dashboard accessibility features
- Self-healing pipeline recovery
- Packaged self-serve deployment option
Most people building monitoring tools are building them for organizations with security teams and budgets. HNIS is built for individuals — people who want to know what’s happening around them, protect what matters to them, and not hand that responsibility to a corporation.
Intelligence should be accessible to everyone, not just those who can afford an enterprise contract.
Built and maintained by goggles616