Novyx Core
Knowledge graph with semantic search
Cryptographically durable knowledge graph for AI agents. Semantic search across versioned artifacts. SHA-256 integrity verification. JSON-LD semantic web standards. Built for research assistants, enterprise memory, and autonomous agent persistence.
Technical Specifications
516 Artifacts Validated
Production-tested knowledge graph with 516 cryptographically signed artifacts. 100% integrity pass rate.
SHA-256 Integrity
Every artifact signed with SHA-256. Immutable audit trail. Tamper-proof verification at query time.
JSON-LD Semantic Search
Full semantic web compliance. Vector embeddings for similarity search. Context-aware knowledge retrieval.
Version Control
Git-like versioning for knowledge graphs. Time-travel queries. Atomic rollback on corruption.
Multi-Tenant Ready
Cryptographic tenant isolation. Zero-knowledge boundaries. Built for SaaS from day one.
Open Core
MIT licensed. Full source access. Community-driven development. No vendor lock-in.
Quick Start
// Install
npm install @novyx/core
// Initialize with cryptographic verification
import { NovyxCore } from '@novyx/core'
const core = new NovyxCore({
integrity: 'sha256',
semantic: true,
})
// Store knowledge with automatic signing
await core.store({
type: 'Artifact',
content: 'Research findings on agent memory...',
metadata: { author: 'user-123', timestamp: Date.now() }
})
// Semantic search across versioned history
const results = await core.search({
query: 'agent memory persistence',
similarity: 0.8,
verifyIntegrity: true
})→ Full docs at github.com/novyxlabs/novyx-core
Built For
Research Assistants
Durable memory for academic research. Version control for evolving hypotheses. Semantic search across literature.
Example: ArXiv agent that remembers every paper it has read
Enterprise Memory
Persistent context for business intelligence. Audit trails for compliance. Multi-tenant knowledge isolation.
Example: CRM agent with complete customer history
Agent Persistence
Memory that survives restarts. Context that compounds over time. Verifiable learning progression.
Example: Autonomous trading agent with years of market memory
Start Building Today
MIT licensed. Production-ready. Built by infrastructure engineers for infrastructure engineers.