NVIDIA Releases Auto Ontology: An Open Semantic Layer That Grounds AI Agent SQL
Reported by MarkTechPost, Hacker News and news.lavx.hu3 sources ↓
NVIDIA released Auto Ontology, an open-source Python framework that builds semantic layers over enterprise data. It reconciles definitions across systems and provides agent tools with context through MCP, CLI and API.
Key points
- Auto Ontology builds enterprise ontologies and knowledge graphs from metadata
- It checks generated queries against certified definitions and returns provenance
- The framework is licensed under Apache 2.0 and is not accepting external contributions
- It can run locally with Docker Compose or on Kubernetes with Helm
- Inference uses NVIDIA NIM endpoints; setup requires an NVIDIA API key
Sources
Details
NVIDIA says Auto Ontology complements existing semantic models by connecting ontologies and reconciling inconsistent definitions across enterprise systems. Its listed functions include discovering metadata, grounding business metrics and relationships, validating generated queries, exposing context to agent harnesses, and incorporating certified corrections into the ontology.
Its default inference setup uses NVIDIA models for reasoning, embeddings and reranking. The project is available under the Apache License 2.0; the supplied materials say it is not currently accepting outside contributions.
Written from coverage as of
Images
- news.lavx.hu
- Hacker News
- Hacker News
- Hacker News
Stocks in this story
Change since open: 20 Nov to 21 Nov 2026
Last refresh