Skip to main content

Introduction to Nexus AI

Nexus AI is the intelligent layer built on top of the Nexus Query Engine. It's a full-fledged AI framework that lets you combine modern large language models, tools, and workflows to build powerful, autonomous agents tailored to your needs.

Nexus works seamlessly with a variety of LLMs, including OpenAI, Google, DeepSeek and even self-hosted models through Ollama. You can configure multiple LLMs at the same time, making it easy to select the right model for the right task.

On top of your configured LLMs, you can create multiple AI Agents. Each agent combines an LLM with a set of tools, all fully MCP-compatible. They empower LLMs to execute complex tasks independently. Nexus integrates with Smithery.ai to give you access to thousands of tools out of the box, and supports tools that require properties (like API keys, credentials, or other configuration) directly through the API.

Nexus comes with three core tools included by default, in addition to partner and external tools:

  1. Metadata lookup — discovers relevant datasets in your virtual data lake based on the agent's prompt.
  2. SQL execution — generates and runs SQL queries using the metadata and interprets the results.
  3. Graph generation — creates visualizations from query results.

All of these tools can be attached to a single AI Agent, which will then autonomously decide which tools to use, in what order, and can even retry steps if something goes wrong; all without human intervention.

Through the API, you can orchestrate workflows that leverage Nexus AI Agents, such as building scheduled tasks that let your agents operate daily, fully automated.

With Nexus, you get an easy-to-work-with, modern framework that supports a broad range of technologies and LLMs, and integrates seamlessly with your data and countless external services, so you can focus on outcomes, not plumbing.