How eidolon-ai/eidolon works
Main Components and Architecture
a) Client Layer:
• Python client (eidolon-ai-client)
• Web UI interface (webui directory)
• REST API endpoints
b) Core Service Layer:
• Agent Service components
• Browser Service
• Usage Service
• Kubernetes Operator
c) SDK Layer:
• Agent Framework
• Processing Units
• Memory Systems
• Security Components
Key Relationships to Highlight
• Client-to-Service communication via HTTP/REST
• Inter-agent communication pathways
• Integration with external services (LLMs, databases)
• Kubernetes deployment flow
Component Details
a) Agent Framework:
┌─ Agent Framework ─────────────┐
• ┌─────────┐ ┌──────────┐ │
• APU │←→ │ Logic │ │
• Units │ │
• ┘ └──────────┘ │
• ┌─────────┐ ┌──────────┐ │
• Memory │←→ │ Security │ │
• Systems │ │ Manager │ │
• ┘ └──────────┘ │
• ┘
b) Service Architecture:
┌─ Services ──────────────────┐
• ┌─────────┐ ┌──────────┐ │
• Agent │←→ │ Browser │ │
• Service │ │ Service │ │
• ┘ └──────────┘ │
• ┌─────────┐ ┌──────────┐ │
• Usage │←→ │ K8s │ │
• Service │ │ Operator │ │
• ┘ └──────────┘ │
• ┘
Elements
• Component labels
• API endpoints
• External service connections
• Security boundaries
• Deployment boundaries
Specific Implementation Details
Integration with:
• LLM providers (OpenAI, Anthropic, etc.)
• Storage systems (MongoDB, S3)
• Authentication systems
• Kubernetes infrastructure
External Integrations
This will provide a clear understanding of how Eidolon's components work together while highlighting its modular and service-oriented nature.