Agentic Context Layer

Multi-agent LangGraph · Trusted data governance
⚡ Live demo

Running the governance pipeline

6 AI agents are analysing a sample e-commerce schema right now — profiling columns, inferring lineage, detecting PII, generating definitions, and scoring trust.

📥
InputParser
Parsing DDL · no LLM · 6 tables, ~32 columns
waiting
🔬
Profiler
gpt-4o-mini · types, nulls, patterns, anomalies
waiting
🕸️
Lineage
gpt-4o-mini · FKs · inferred deps · orphans
waiting
🔒
PII Detector
Deterministic · 9 categories · DDL masking before LLM
waiting
📖
Semantic
gpt-4o · business definitions + domain + semantic typing
waiting
🎯
Trust Scorer
gpt-4o + deterministic rules · calibrated 0–1 per definition
waiting
📦
Aggregator
Assembling governed context layer · flushing audit trail
waiting
Cold start: the first request spins up the LLM calls fresh — this usually takes 25–35 s. Results are cached so every subsequent visit is instant.
Elapsed: 0s

Pipeline failed

Something went wrong while running the agents. Check that OPENAI_API_KEY is set on the server.