📂 Data Categories

Control what types of data are allowed into the knowledge graph.

Category	Behavior	Description
👤 Personal data		GDPR Art. 6 — requires legal basis
💰 Financial data		Banking, transactions — encrypted at rest
🏥 Health data		GDPR Art. 9 special category — denied by default
🏢 Company data		Business information
🌐 Public data		Publicly available data

⚙️ Quality Gates

Measurable rules for data quality before storage.

Min relevance for storage 🔒 0.7 Below this threshold → not stored

Max duplicate similarity 0.95 Above this → rejected as duplicate

Max entry size (tokens) 🔒 2000 Larger records are truncated or rejected

TTL (days) Auto-expire records after this many days

Min record length (chars) Shorter records are rejected

Rate limit (writes/min) Max ingestion rate

Required classification — every record must have ≥1 category tag PII detection — auto-scan for emails, phone numbers, IDs