Sovereign Watch v0.10.3: Distributed Multi-INT Fusion Center

Operational Status: Phase 2 (Tactical Intelligence & Tracking) - Active Development

Sovereign Watch is a self-hosted, distributed intelligence fusion platform designed to ingest, normalize, and analyze high-velocity telemetry (ADS-B, AIS, Orbital) and high-variety intelligence (SIGINT, OSINT). It enforces data sovereignty by running entirely on local hardware (Edge to Cloud), utilizing a "Pulse" architecture for data collection and a "Tiered AI" strategy for cognition.

---

---

🛠️ Quick Start

Prerequisites

• Docker & Docker Compose
• NVIDIA Container Toolkit (if using Local AI/Jetson)

Installation

1. Clone & Configure:

   cp .env.example .env
   # Edit .env with your keys & config:
   # - CENTER_LAT / CENTER_LON (Your monitoring area)
   # - AISSTREAM_API_KEY (Maritime feed)
   # - ANTHROPIC_API_KEY / GEMINI_API_KEY (LLM Cognition)
   # - VITE_MAPBOX_TOKEN (3D Terrain & Maps)
   # - KIWI_HOST / KIWI_PORT (JS8Call SDR source)
   # - MY_GRID (Your Maidenhead locator)

2. Boot System:

   docker compose up -d --build

3. Access Interfaces:

   • Tactical Map (UI): http://localhost:3000
   • Fusion API: http://localhost:8000/docs
   • Redpanda Console: http://localhost:8080

⚠️ Disclaimer & Liability

📡 Source Data and Open Intelligence

Sovereign Watch ingests telemetry and intelligence from public, open-source networks (e.g., ADS-B, AIS, public API feeds). The positional data, classifications, and intelligence displayed within this platform are strictly derivative of these unencrypted, publicly broadcasted signals.

🛡️ Limited Liability

All data is provided "AS IS" without any warranty of accuracy, reliability, or completeness. The developers and maintainers of Sovereign Watch assume no responsibility or liability for:

• The accuracy of real-time or historical tracking information.
• Decisions or actions taken based on the intelligence presented by this software.
• Disruptions to the third-party networks providing the upstream data.

Sovereign Watch is designed purely for research, educational, and hobbyist data fusion purposes.

---

📂 Architecture Overview

graph TD
    subgraph "Ingestion (Python Pollers)"
        A[ADS-B Network] -->|JSON| B(Ingestion Services)
        C[AIS Stream] -->|JSON| B
        Z[Orbital TLE Feed] -->|TLE| B
        J[JS8Call HF Radio] -->|TCP API| B
        B -->|TAK Protobuf| D(Redpanda Bus)
    end

    subgraph "Persistence (TimescaleDB)"
        D -->|Stream| E[(Tracks Hypertable)]
        D -->|Stream| F[(Vector Store)]
    end

    subgraph "Cognition (LiteLLM)"
        G[Fusion API] -->|Query| H{AI Router}
        H -->|Tier 1| I[Local Llama3]
        H -->|Tier 3| J[Claude 3.5]
    end

    subgraph "Presentation (React + Deck.gl)"
        K[MainHUD Shell] --> L[Intelligence Feed]
        K --> M[Projective Velocity Blending]
        M -->|WebGL 3D| N[Mapbox / CARTO Overlay]
        K --> O[Radio Terminal]
    end

Loading

🛡️ Tactical Design ("Sovereign Glass")

• Chevron-First Architecture: Unified directional trackers for all assets; no legacy dot markers.
• Hybrid 3D Engine: Seamlessly switches between Mapbox 3D (Terrain/Satellite) and CARTO Dark Matter (Vector/Local) based on configuration.
• High-Fidelity HUD: Integrated global TopBar with synchronized temporal references (UTC), real-time entity tracking sidebars, and active intelligence feeds.
• Immersion Layers: Micro-noise texture and tactical grid overlays for a professional surveillance aesthetic.
• Interactive Vectors: Pickable chevrons for target locking, historic trail inspection, entity telemtry drill-down, and tactical time travel (replay).

🗼 Tactical Indicators

Asset Symbology

• Chevrons: Indicate directional heading and asset type (Aviation/Maritime). Hovering/Clicking reveals the target's specific classification.
• ⊕ Cross/Star: Orbital assets (Satellites). Rendered at ground-track position with predicted orbital paths.
• Pulsating Rings: Active telemetry updates. Intensity increases when an asset is selected.
• Tactical Outline: High-value/special assets (SAR, Military, Law Enforcement vessels, Drones, Helicopters) emit a glowing Tactical Orange signature aura for instantaneous operator recognition.

Intelligent Color Coding

The Tactical Map uses dynamic "thermal" gradients to visualize critical metadata:

Aviation (Altitude)

• 🟢 Green: Grounded / Low (< 5,000ft)
• 🟡 Yellow: Lower-Altitude / Approach (~ 10,000ft)
• 🟠 Orange: Mid-Altitude Climb/Descent (~ 20,000ft)
• 🔴 Red: High-Altitude Cruise (~ 30,000ft)
• 🟣 Magenta: Very High-Altitude (> 40,000ft)

Maritime (Speed)

• 🔵 Dark Blue: Stationary / Anchored (0 kts)
• 🟦 Medium Blue: Harbor Speed / Patrolling (< 10 kts)
• 🩵 Light Blue: Cruising (~ 15 kts)
• ⚪ Cyan/White: High-Speed Transit (25+ kts)

Orbital (Category)

• 💎 Sky Blue: GPS & Navigation Constellations
• 🟠 Amber: Weather & Environmental Monitoring
• 🟢 Emerald: Communication & Internet (Starlink/OneWeb)
• 🔴 Rose: Surveillance & Known ISR Satellites
• ⚪ Gray: Other / Unclassified Satellites

🔍 Core Capabilities

• Deep Vessel Classification: Real-time parsing of Maritime ShipStaticData to classify tankers, cargo, military, SAR, and passenger vessels with absolute precision.
• Orbital Pulse Tracking: End-to-end satellite tracking using Celestrak TLE ingestion and live SGP4 propagation (accuracy updated every 30s) to visualize LEO/MEO/GEO assets.
• JS8Call Signal Intelligence: Integrated HF digital mode (JS8) radio bridge and interactive HUD terminal for real-time tactical communications and station tracking.
• Projective Velocity Blending (PVB): Physics-based kinematic rendering ensures fast-moving aircraft coast smoothly between delayed transponder pings, with zero "rubber-banding."
• Granular Filtering Matrix: Advanced HUD tools to strip away visual noise. Filter the theater by specific sub-classes (e.g., hiding generic cargo and passenger jets, while highlighting Drones, Helicopters, and Military fast-movers).
• Time-Travel (Historian Service): All positional data is written to a TimescaleDB instance. Operators can search for past targets and "replay" tactical situations from hours or days ago directly within the WebGL interface.

📂 Directory Structure

Path	Purpose	Git Status
/AGENTS.md	Master Guide for AI Developers (Read This First)	Tracked
/.agent	Agent memory, skills, and global project rules.	Tracked
/backend/ingestion	Python multi-source polling frameworks.	Tracked
/backend/db	Database schema (init.sql) and migration scripts.	Tracked
/backend/api	Python FastAPI service for Fusion and Analysis.	Tracked
/js8call	JS8Call HF Radio Terminal container and bridge.	Tracked
/frontend	React + Vite application (Tactical Map + HUD).	Tracked
/docs	Architecture plans, research, and progress logs.	Tracked

🤖 AI Agent Protocol

This repository is Agent-Aware. If you are an AI assistant contributing to this project:

1. Read Rules: You MUST read AGENTS.md at the start of your session. It is the authoritative entry point.
2. Environment Protocol: Never run commands (npm, pip, python) directly on the host. Always use the Docker Compose commands defined in the rules.
3. Communication: All inter-service data must adhere to the TAK Protocol (Protobuf) as defined in tak.proto.
4. Aesthetics: Follow the "Sovereign Glass" design principles for all UI modifications.

🧪 Development Workflow

🐳 The "Container-First" Rule

Never run commands (npm, node, python, pip, etc.) directly on the host. ALL interactions and execution must happen through Docker Compose.

• Starting Services: docker compose up -d (or docker compose up -d --build <service> after dependency changes)
• Running One-off Tasks: docker compose run --rm <service> <command>
• Viewing Logs: docker compose logs -f <service>

⚡ Live Updates (HMR)

Both Frontend and Backend services are configured for Hot Module Replacement:

• Frontend: Save any .tsx/.ts/.css file. Vite automatically syncs changes instantly (polling, 1s interval). No restart required.
• Backend: Save any .py file. Uvicorn reloads automatically. No restart required.
• Ingestion/Misc Services: Sometimes require restarts (docker compose restart <service>) upon configuration changes.

Note: Only rebuild containers when altering Dockerfile configurations or modifying dependencies.

---

Maintained by d3FRAG Networks & The Antigravity Agent Team.