Skip to content

quickstart.mdx

Latest commit

 

History

History
75 lines (60 loc) · 2.51 KB

quickstart.mdx

File metadata and controls

75 lines (60 loc) · 2.51 KB
Raw
title NVIDIA NeMo Fabric Quickstart
description Get started with NVIDIA NeMo Fabric.
template-library-version 1.0.0

{/* SPDX-FileCopyrightText: Copyright (c) 2026, NVIDIA CORPORATION & AFFILIATES. All rights reserved. SPDX-License-Identifier: Apache-2.0 */}

This guide demonstrates how to install and run a simple agent using NeMo Fabric.

Install NeMo Fabric

Hermes Agent 0.20 and later is no longer installable from PyPI. Follow the Hermes Agent installation guide, then install NeMo Fabric and its bare Hermes adapter into the Python environment that runs Hermes Agent:

python -m pip install nemo-fabric nemo-fabric-adapters-hermes

These packages do not install Hermes Agent. For local development from a NeMo Fabric checkout, run just install-hermes-agent from the repository root instead. The recipe checks out tested Hermes source with Relay 0.10 support and installs it in an isolated environment. Set ADAPTER_PYTHON to .venv-hermes/bin/python when running NeMo Fabric from the project environment. Refer to the NeMo Fabric installation guide for separate adapter environments and other supported package combinations.

Obtaining API Keys

The following code example uses an NVIDIA-hosted model and requires an NVIDIA API key defined with the NVIDIA_API_KEY environment variable. An API key can be obtained by creating an account on build.nvidia.com.

Run a Simple Agent

import asyncio

from nemo_fabric import (
    Fabric,
    FabricConfig,
    HarnessConfig,
    MetadataConfig,
    ModelConfig,
    RuntimeConfig,
)

config = FabricConfig(
    metadata=MetadataConfig(name="quickstart-agent"),
    harness=HarnessConfig(
        adapter_id="nvidia.fabric.hermes",
        resolution="preinstalled",
    ),
    runtime=RuntimeConfig(max_turns=1),
    models={
        "default": ModelConfig(
            provider="nvidia",
            model="nvidia/nemotron-3-nano-omni-30b-a3b-reasoning",
            api_key_env="NVIDIA_API_KEY",
            base_url="https://integrate.api.nvidia.com/v1",
        )
    },
)

result = asyncio.run(Fabric().run(config, input="Who are you?"))
print(result.output.response)

A more detailed version of this example is available as a Jupyter Notebook at examples/notebooks/01_quickstart.ipynb.

Next Steps