The agent examples demonstrate how NeMo Agent Toolkit accelerates and enables AI Agent development.
The examples showcase 5 distinct AI Agent architectures solving a similar problem in different ways.
By leveraging the NeMo Agent Toolkit plugin system and the Builder object, you can use both pre-built and
custom agentic workflows and tools in a flexible manner.
If you have not already done so, follow the instructions in the Install Guide to create the development environment and install NeMo Agent Toolkit.
If you have not already done so, follow the Obtaining API Keys instructions to obtain an NVIDIA API key. You need to set your NVIDIA API key as an environment variable to access NVIDIA AI services:
export NVIDIA_API_KEY=<YOUR_API_KEY>Each agent example contains its own installation and usage instructions. Navigate to the specific example directory and follow the README instructions:
- ReAct Agent: See react/README.md for step-by-step reasoning agent implementation
- Tool Calling Agent: See tool_calling/README.md for direct tool invocation agent
- Mixture of Agents: See mixture_of_agents/README.md for multi-agent orchestration
- ReWOO Agent: See rewoo/README.md for planning-based agent workflow
- Auto Memory Wrapper Agent: See auto_memory_wrapper/README.md for the automatic memory wrapper agent workflow
For a deeper dive into the AI Agents utilized in the examples, refer to the component documentation: