NVIDIA NeMo Agent Toolkit Overview — NVIDIA NeMo Agent Toolkit (1.8)
NVIDIA NeMo Agent Toolkit is a flexible, lightweight, and unifying library that allows you to easily connect existing enterprise
to data sources and
across any framework.
Install
uv
uvpipinstallnvidia-nat pip
pipinstallnvidia-nat For detailed installation instructions, including optional dependencies, please refer to the
.
Key Features
NeMo Agent Toolkit works side-by-side and around existing agentic frameworks, such as
,
,
,
,
, as well as customer enterprise frameworks and simple Python agents. This allows you to use your current technology stack without replatforming. NeMo Agent Toolkit complements any existing agentic framework or
tool you’re using and isn’t tied to any specific agentic framework, long-term memory, or data source.
Every agent, tool, and agentic
in this library exists as a function call that works together in complex software applications. The composability between these agents, tools, and workflows allows you to build once and reuse in different scenarios.
Start with a pre-built agent, tool, or workflow, and customize it to your needs. This allows you and your development teams to move quickly if you’re already developing with agents.
Use the profiler to profile entire workflows down to the tool and agent level, track input/output tokens and timings, and identify bottlenecks.
Monitor and debug your workflows with dedicated integrations for popular observability platforms such as LangSmith, Phoenix, Weave, and Langfuse, plus compatibility with OpenTelemetry-based systems. Track performance, trace execution flows, and gain insights into your agent behaviors.
Validate and maintain accuracy of agentic workflows with built-in evaluation tools.
Use the NeMo Agent Toolkit UI chat interface to interact with your agents, visualize output, and debug workflows.
Compatible with
. You can use NeMo Agent Toolkit as an
to connect to and use tools served by remote MCP servers. You can also publish tools with the
runtime or an MCP server using the
.
Compatible with
. You can use NeMo Agent Toolkit as an
to connect to and delegate tasks to remote A2A agents. You can also use NeMo Agent Toolkit as an
to publish workflows as discoverable A2A agents.
Hello World Example
Before getting started, it’s possible to run this simple workflow and many other examples in Google Colab with no setup. Click here to open the introduction notebook:
.
Install NeMo Agent Toolkit along with the LangChain integration plugin:
uv
uvpipinstall"nvidia-nat[langchain]" pip
pipinstall"nvidia-nat[langchain]"
Ensure you have set the NVIDIA_API_KEY environment variable to allow the example to use NVIDIA NIMs. An API key can be obtained by visiting
and creating an account.
exportNVIDIA_API_KEY=<your_api_key>
Create the NeMo Agent Toolkit workflow configuration file. This file will define the agents, tools, and workflows that will be used in the example. Save the following as workflow.yml:
functions:# Add a tool to search wikipediawikipedia_search:_type:wiki_searchmax_results:2llms:# Tell NeMo Agent Toolkit which LLM to use for the agentnim_llm:_type:nimmodel_name:meta/llama-3.1-70b-instructtemperature:0.0workflow:# Use an agent that 'reasons' and 'acts'_type:react_agent# Give it access to our wikipedia search tooltool_names:[wikipedia_search]# Tell it which LLM to usellm_name:nim_llm# Make it verboseverbose:true# Retry up to 3 timesparse_agent_response_max_retries:3
Run the Hello World example using the nat CLI and the workflow.yml file.
natrun--config_fileworkflow.yml--input"List five subspecies of Aardvarks"This will run the workflow and output the results to the console.
Workflow Result:['Here are five subspecies of Aardvarks:\n\n1. Orycteropus afer afer (Southern aardvark)\n2. O. a. adametzi Grote, 1921 (Western aardvark)\n3. O. a. aethiopicus Sundevall, 1843\n4. O. a. angolensis Zukowsky & Haltenorth, 1957\n5. O. a. erikssoni Lönnberg, 1906']
FAQs
For frequently asked questions, refer to
.
Feedback
We would love to hear from you! Please file an issue on
if you have any feedback or feature requests.