Capsule Architecture
Knowledge Capsules are structured files stored within AMX Workspaces, providing encapsulated units of knowledge that can be versioned, shared, and processed by AMX Engine. The architecture treats capsules as files within the workspace’s knowledge/ directory, managed through git-like operations.
Workspace Integration
Knowledge Capsules exist as files within the workspace structure:
workspace/
├── .aimatrix/
├── knowledge/
│ ├── capsules/
│ │ ├── customer-service.kc # Customer service capsule
│ │ ├── product-catalog.kc # Product knowledge capsule
│ │ └── troubleshooting.kc # Support knowledge capsule
│ ├── volumes/
│ └── libraries/
├── agents/
├── workflows/
└── models/File-Based Management
- Storage: Capsules are
.kcfiles stored inknowledge/capsules/ - Version Control: Managed through AMX Hub like git repositories
- Processing: AMX Engine reads capsule files directly from workspace
- Collaboration: Shared through workspace push/pull operations
System Architecture Overview
graph TB
subgraph "External Layer"
DEV[Developer Portal]
MARKET[Marketplace UI]
API[External APIs]
end
subgraph "API Gateway"
GATEWAY[Authentication & Routing]
end
subgraph "Core Services"
REGISTRY[Capsule Registry]
LIFECYCLE[Lifecycle Manager]
DEPENDENCY[Dependency Resolver]
SANDBOX[Sandboxing Engine]
end
subgraph "Data Layer"
SUPABASE[(Supabase)]
VECTOR[(Vector Store)]
EVENTS[(Event Store)]
CACHE[(Redis Cache)]
end
subgraph "Runtime Environment"
INSTANCES[Capsule Instances]
ISOLATION[Isolation Layer]
MONITORING[Resource Monitor]
end
DEV --> GATEWAY
MARKET --> GATEWAY
API --> GATEWAY
GATEWAY --> REGISTRY
GATEWAY --> LIFECYCLE
REGISTRY --> DEPENDENCY
LIFECYCLE --> SANDBOX
DEPENDENCY --> SANDBOX
SANDBOX --> INSTANCES
INSTANCES --> ISOLATION
ISOLATION --> MONITORING
REGISTRY --> SUPABASE
LIFECYCLE --> EVENTS
SANDBOX --> VECTOR
MONITORING --> CACHE
Capsule Structure and Format
Capsule Manifest
Every knowledge capsule is defined by a comprehensive manifest file that describes its structure, dependencies, and capabilities.
|
|
Directory Structure
knowledge-capsule/
├── manifest.json # Capsule manifest
├── README.md # Documentation
├── CHANGELOG.md # Version history
├── LICENSE # License file
├── src/ # Source code
│ ├── handlers/ # Event handlers
│ ├── endpoints/ # API endpoints
│ ├── processors/ # Data processors
│ └── utilities/ # Helper functions
├── knowledge/ # Knowledge assets
│ ├── schemas/ # Data schemas
│ ├── rules/ # Business rules
│ ├── workflows/ # Process definitions
│ └── embeddings/ # Vector embeddings
├── tests/ # Test suite
│ ├── unit/ # Unit tests
│ ├── integration/ # Integration tests
│ └── fixtures/ # Test data
├── docs/ # Additional documentation
└── examples/ # Usage examplesDependency Management System
Dependency Resolution Algorithm
The dependency resolver uses a sophisticated algorithm to ensure compatibility and prevent conflicts:
|
|
Version Compatibility
The system uses semantic versioning with advanced constraint resolution:
|
|
Isolation and Sandboxing
Runtime Sandboxing
Each capsule runs in an isolated environment to prevent interference and security issues:
|
|
Permission System
|
|
Supabase Integration
Database Schema
The complete database schema for the Knowledge Capsule System:
|
|
Event Sourcing Implementation
|
|
Performance Optimization
Caching Strategy
|
|
Security Considerations
Code Scanning and Validation
|
|
This architecture provides a robust foundation for the Knowledge Capsule System, ensuring security, scalability, and maintainability while enabling powerful extensibility for AIMatrix users.