Architecture

CredB Nexus is planned as an offline-first distributed file indexing and retrieval platform for local networks.

Core Components

• crawler agents
• backend
• frontend
• metadata database
• optional local LLM layer

Rules

• crawling stays separate from LLM execution
• deterministic metadata search must work without semantic search
• semantic search is an enhancement, not the base retrieval path
• deployment should remain offline-friendly