What is NVIDIA Auto Ontology?

NVIDIA Auto Ontology is an open library designed to build a context layer from enterprise metadata, reconcile definitions across systems, and expose it to agent harnesses operación histórica respaldada por Nvidia. It allows developers to discover enterprise ontologies, ground agents with trusted business meanings, validate queries against certified definitions, and expose context via MCP, CLI, and API.

The project is distributed under the Apache License 2.0 and integrates with herramienta de código abierto viral ecosystems to enhance reasoning layers over structured data. It also complements existing semantic models and supports text-to-SQL benchmarks.

What are the core software components and architecture?

The stack is composed of a Next.js 16 frontend with Better Auth, a FastAPI backend in Python 3.12+ for chat and text-to-SQL translation, an ingestion worker, and Postgres with pgvector nueva herramienta open-weight. Optional HashiCorp Vault is supported for secure credential storage.

Models are configured using environment variable triplets such as endpoint, API key, and model name, defaulting to NVIDIA NIM endpoints and models like Nemotron.

How do you install and deploy Auto Ontology locally?

To set up the stack locally using Docker Compose, ensure you have Docker with Docker Compose v2 and an active NVIDIA API key.

Follow these steps to complete the installation and startup process:

  1. Clone the repository using git clone https://github.com/NVIDIA/auto-ontology auto_ontology and change into the directory.
  2. Create your environment configuration file by copying cp.env.example.env.
  3. Edit the.env file to fill in your Postgres credentials, AUTH_SECRET, APP_URL, and DEFAULT_MODELS_API_KEY.
  4. Build the images and start the services by executing docker compose up -d --build.
  5. Access the web user interface at http://localhost:3000 to manage connections and run queries.