The AiMatrix initiative aims to deliver a unified, extensible platform that brings together advanced AI-driven multi-agent systems, digital twins, and seamless enterprise integration. At its core, AiMatrix provides a scalable hub for orchestrating autonomous agents—ranging from cloud-hosted services to on-device daemons—while leveraging the proven BigLedger API backbone for data storage, transaction handling, and auditability. This architecture enables organizations to model complex processes, simulate real-world systems, and execute automated workflows with minimal overhead.
As depicted in the high-level architecture diagram on page 1 of the attached document, AiMatrix is structured around three primary layers:
AiMatrix Apps (Mobile & Desktop): Kotlin-based MCP clients with embedded SQLite offline databases allow agents to operate reliably even when disconnected.
AiMatrix Daemons & CLI Tools: Background monitoring and auto-update agents written in Kotlin and Python deliver seamless setup assistance and remote management capabilities.
AiMatrix Console: A unified command-line interface for administrators and developers to deploy models, manage registries, and interact with agents.
MCP Server / Gateway (xxx.aimatrix.net): Acts as the central routing fabric, built on Spring Cloud Gateway, Kotlin KTOR, and Python FastAPI microservices.
Global Directory & Registry: Maintains metadata about all remote agents, sensors, and “coworkers,” enabling dynamic discovery and secure communication.
Planner & Executor: Implements a flexible planning engine and task scheduler, coordinating multi-agent workflows and digital twin simulations (e.g., via Kalasim).
Shared & Individual Models: Supports both isolated and collaborative AI models, including RAG-powered knowledge graphs and digital twin replicas for real-time state tracking.
BigLedger API (Prod/Dev): Provides reliable, versioned access to ledger features (transactions, document storage, audit trails) via DynamoDB/S3 backends.
AiLedger & MongoDB: Handles high-throughput event ingestion and schema-flexible data storage for AI training pipelines and real-time analytics.
ETL Marketplace Connectors: Prebuilt integrations with e-commerce platforms (Lazada, Shopee, TikTok), payment gateways (iPay88, GDex, RHB), and legacy POS/ERP systems (SAP, Anchanto).
External Services & Communication Channels: Includes signing cloud for eKYC/SSM, government tax interfaces, and chat-bot connectors for Messenger, Telegram, and WeChat.
Underpinning these layers is a comprehensive set of Kotlin and Python libraries for linear algebra, multi-agent programming, simulation, planning, and UI components. This modular toolkit empowers developers to rapidly prototype custom digital twins, embed intelligent behaviors, and integrate with corporate portals or consumer-facing applications.
By combining offline-capable clients, a robust cloud hub, and a rich ecosystem of integrations, AiMatrix delivers an end-to-end solution for organizations seeking to harness AI for process automation, complex system modeling, and real-time decision support—whether in manufacturing, supply chain, or service operations.