Vector Search and AI Integration

Vektagraf supports vector fields:

"embedding": {
  "type": "vector",
  "dimensions": 1536,
  "index": "hnsw"
}

Motivation:

• AI and LLM-based systems require semantic search
• Vector indexes must integrate with tenant boundaries
• Vector fields must participate in provenance
• Embeddings must be validated at ingestion

This makes Vektagraf AI-native.