PROJECT NEXUS
The Global Neural Hub
The centralized island architecture connecting 72 subsystems across 9 platforms through a three-pillar database convergence, enriching every AI request with context drawn from 61,271 memory parameters before routing to any of three global LLM providers.
Overview
The Convergence Point
Project Nexus is the centralized island architecture at the heart of the DragonSource ecosystem. Hosted on Google Cloud Platform and launched publicly on April 1, 2026, it serves as the single convergence point for every application, subsystem, database, and AI agent across the entire network — 72 subsystems spanning 9 production platforms, unified under one operational core.
At its center is the Tower — the API exchange core where a three-pillar database architecture (Neon × Supabase × Firebase) merges structured relational data, real-time telemetry, and document-based memory into a single contextualized payload. Before any request reaches an LLM endpoint, it passes through 61,271 keyword-destination pairs drawn from 78 memory parameter files, transforming generic artificial intelligence into deeply personalized, context-aware intelligence.
This is not a microservices mesh or an API gateway. This is a neural hub — a topographical island where data flows converge, context is forged, and every system in the ecosystem speaks through a single, authorized voice. Nexus is the infrastructure layer that makes the DragonSource vision possible: AI that knows you, understands context, and operates with institutional memory.
Telemetry
By the Numbers
A comprehensive inventory of every subsystem, memory parameter, AI agent, and database layer that flows through the Neural Hub architecture.
Foundation
Three Database Pillars
The geological layers beneath the Island — the bedrock the Storage Containers and Tower are built upon. When a request arrives, the Tower performs a three-pillar fetch, merging all streams into a single contextualized packet.
Neon
PostgreSQL
The structured relational backbone. Platform configurations, user management, governance rules, and operational state — queryable, transactional, and strongly typed.
Feeds → Provides the operational context layer for every Tower request — platform state, session data, and configuration parameters.
Supabase
Real-time Layer
The nervous system. Real-time data streams, authentication, live telemetry, and event-driven triggers that keep every subsystem synchronized.
Feeds → Delivers live state and authentication context — ensuring every AI response reflects the current moment.
Firebase
Document Store
The document-centric memory layer. Firestore collections, real-time sync, cloud functions, and the flexible schema that adapts to the evolving ecosystem.
Feeds → Supplies the personal memory and Mind Map data — 61,271 keyword-destination pairs for personalized intelligence.
Deployment Sequence
The Six Operations
Six strategic phases that will carry the DragonSource ecosystem from launch to global deployment. Each operation is a self-contained mission with distinct objectives, timelines, and outcomes.
The Sabbatical
EVA's global doctoral residency
The Campaign
Strategic deployment across target markets
The Evolution
Systematic capability upgrades
The Architect
Infrastructure scaling & refinement
The Crucible
Stress testing & adversarial hardening
The Syndicate
Partnership & ecosystem expansion
Infrastructure
What Powers the Hub
The production stack behind Nexus — from cloud hosting and serverless compute to the three database pillars and three LLM providers that power every AI interaction.
GCP
Google Cloud Platform
Hosting & Compute
RUNTIME
Cloud Run
Serverless Containers
PILLAR 01
Neon PostgreSQL
Relational Layer
PILLAR 02
Supabase
Real-time & Auth
PILLAR 03
Firebase
Document Store & Functions
LLM
OpenAI
GPT-4o / o1 / o3
LLM
Google Gemini
Gemini 2.5 Pro
LLM
Anthropic
Claude Sonnet 4
Global Network
The Operational Map
22 sovereign hub nodes are live across 6 continents, with 17 more under evaluation for Wave 2 deployment.
Active Nodes
22 Hubs • 86 Districts
The full operational directory — every hub commander, district, and target across the live global neural network.
Wave 2 Expansion
17 Planned • 68 Districts
The next wave — 17 hub nodes under evaluation across strategic corridors including DC, Boston, Seattle, Berlin, Taipei, and Tel Aviv.