Frameworks Overview#

NVIDIA NeMo Agent toolkit (NAT) provides comprehensive support for multiple agentic frameworks, allowing you to use your preferred development tools while leveraging the capabilities of NAT. This document describes the framework integrations available and their respective levels of support.

Supported Frameworks#

NeMo Agent toolkit integrates with the following frameworks:

ADK: Google Agent Development Kit for building AI agents
Agno: A lightweight framework for building AI agents
CrewAI: A framework for orchestrating role-playing AI agents
LangChain/LangGraph: A framework for developing applications powered by large language models
LlamaIndex: A data framework for building LLM applications
Semantic Kernel: Microsoft’s SDK for integrating LLMs with conventional programming languages

Framework Support Levels#

NeMo Agent toolkit provides different levels of support for each framework across the following dimensions:

LLM Provider Support#

The ability to use various large language model providers with a framework, including NVIDIA NIM, OpenAI, Azure OpenAI, AWS Bedrock, and LiteLLM.

Embedder Provider Support#

The ability to use embedding model providers for vector representations, including NVIDIA NIM embeddings, OpenAI embeddings, and Azure OpenAI embeddings.

Retriever Provider Support#

The ability to integrate with vector databases and retrieval systems, such as NeMo Retriever and Milvus.

Tool Calling Support#

The ability to use framework-specific tool calling mechanisms, allowing agents to invoke functions and tools during execution.

Profiling Support#

The ability to view workflow execution traces including intermediate steps, LLM calls, and tool calls within the NeMo Agent toolkit profiler.

Framework Capabilities Matrix#

The following table summarizes the current support level for each framework:

Framework	LLM Providers	Embedder Providers	Retriever Providers	Tool Calling	Profiling
ADK	✅ Yes	❌ No	❌ No	✅ Yes	✅ Yes
Agno	⚠️ Limited	❌ No	❌ No	✅ Yes	✅ Yes
CrewAI	✅ Yes	❌ No	❌ No	✅ Yes	✅ Yes
LangChain	✅ Yes	✅ Yes	✅ Yes	✅ Yes	✅ Yes
LlamaIndex	✅ Yes	✅ Yes	❌ No	✅ Yes	✅ Yes
Semantic Kernel	⚠️ Limited	❌ No	❌ No	✅ Yes	✅ Yes

Framework-Specific Details#

ADK (Google Agent Development Kit)#

Google’s Agent Development Kit (ADK) is a framework for building AI agents with multiple LLM providers. It provides a set of tools for creating agents that can be used to create complex workflows powered by LLMs. ADK focuses on modularity and extensibility, making it suitable for integrating custom data pipelines and enhancing intelligent applications.

For more information, visit the ADK website.

Capability	Providers / Details
LLM Providers	NVIDIA NIM, OpenAI, Azure OpenAI, AWS Bedrock, LiteLLM
Embedder Providers	None (use framework-agnostic embedders if needed)
Retriever Providers	None (use ADK native tools)
Tool Calling	Fully supported through the ADK `FunctionTool` interface
Profiling	Comprehensive profiling support with instrumentation

Installation:

uv pip install "nvidia-nat[adk]"

Agno#

Agno is a lightweight framework for building AI agents. It provides a set of tools for creating agents that can be used to create complex workflows powered by LLMs. Agno focuses on modularity and extensibility, making it suitable for integrating custom data pipelines and enhancing intelligent applications.

For more information, visit the Agno website.

Capability	Providers / Details
LLM Providers	NVIDIA NIM, OpenAI, LiteLLM
Embedder Providers	None (use framework-agnostic embedders if needed)
Retriever Providers	None (use Agno native tools)
Tool Calling	Fully supported through Agno’s tool interface
Profiling	Comprehensive profiling support with instrumentation

Installation:

uv pip install "nvidia-nat[agno]"

CrewAI#

CrewAI is a framework designed for orchestrating teams of role-playing AI agents that can collaborate and complete complex tasks. It enables the creation of agents with distinct roles, goals, and tools, allowing for multi-agent workflows adaptable to a wide range of scenarios—from research assistants to business process automation.

For more information, visit the CrewAI website.

Capability	Providers / Details
LLM Providers	NVIDIA NIM, OpenAI, Azure OpenAI, AWS Bedrock, LiteLLM
Embedder Providers	None (use framework-agnostic embedders if needed)
Retriever Providers	None (use CrewAI native tools)
Tool Calling	Fully supported through CrewAI’s tool system
Profiling	Comprehensive profiling support with instrumentation

Installation:

uv pip install "nvidia-nat[crewai]"

LangChain/LangGraph#

LangChain is a framework for building applications that utilize large language models (LLMs) to interact with data. It provides a set of tools for creating chains of LLM calls, allowing for complex workflows powered by LLMs. LangChain focuses on modularity and extensibility, making it suitable for integrating custom data pipelines and enhancing intelligent applications.

For more information, visit the LangChain website.

Capability	Providers / Details
LLM Providers	NVIDIA NIM, OpenAI, Azure OpenAI, AWS Bedrock, LiteLLM
Embedder Providers	NVIDIA NIM, OpenAI, Azure OpenAI
Retriever Providers	NeMo Retriever, Milvus
Tool Calling	Fully supported through LangChain’s `StructuredTool` interface
Profiling	Comprehensive profiling support with callback handlers

Installation:

uv pip install "nvidia-nat[langchain]"

LlamaIndex#

LlamaIndex is a powerful framework for building applications that utilize large language models (LLMs) to query and interact with structured and unstructured data. It provides a set of tools for creating indexes over data sources—such as documents, databases, and APIs—enabling complex retrieval, question answering, and orchestration workflows powered by LLMs. LlamaIndex focuses on modularity and extensibility, making it suitable for integrating custom data pipelines and enhancing intelligent applications.

For more information, visit the LlamaIndex website.

Capability	Providers / Details
LLM Providers	NVIDIA NIM, OpenAI, Azure OpenAI, AWS Bedrock, LiteLLM
Embedder Providers	NVIDIA NIM, OpenAI, Azure OpenAI
Retriever Providers	None (Use LlamaIndex native retrievers)
Tool Calling	Fully supported through LlamaIndex’s `FunctionTool` interface
Profiling	Comprehensive profiling support with callback handlers

Installation:

uv pip install "nvidia-nat[llama-index]"

Semantic Kernel#

Microsoft’s Semantic Kernel is a framework for building applications that utilize large language models (LLMs) to interact with data. It provides a set of tools for creating kernels that can be used to create complex workflows powered by LLMs. Semantic Kernel focuses on modularity and extensibility, making it suitable for integrating custom data pipelines and enhancing intelligent applications.

For more information, visit the Semantic Kernel website.

Capability	Providers / Details
LLM Providers	OpenAI, Azure OpenAI
Embedder Providers	None (use framework-agnostic embedders if needed)
Retriever Providers	None (use Semantic Kernel native connectors)
Tool Calling	Fully supported through Semantic Kernel’s function calling
Profiling	Comprehensive profiling support with instrumentation

Installation:

uv pip install "nvidia-nat[semantic-kernel]"