Agent Node on MindPal Multi-Agent Workflow
The Agent Node is the most used node on MindPal Multi-Agent Workflow, and it's a crucial component in building a workflow that automates complex tasks. In this documentation, we'll explore the Agent Node in detail and provide you with the information you need to get started.
What is an Agent Node?
An Agent Node consists of two essential elements:
- The Agent to Use: This is the AI agent that will perform a specific task within the workflow.
- The Instruction: This is the input or prompt that you provide to the agent, guiding it to produce the desired output.
Referencing Data from Other Nodes
One of the powerful features of the Agent Node is the ability to reference data from other nodes in the workflow. You can do this using the @ or # syntax:
- @: Use the
@symbol to refer to a Human Input Node. For example, if you have a Human Input Node that collects a blog post title, you can reference it in the Agent Node instruction using@blog_post_title. - #: Use the
#symbol to refer to the result from a previous node. For example, if you have a node that generates a list of keywords, you can reference it in the Agent Node instruction using#keywords.
By referencing data from other nodes, you can create a dynamic and interactive workflow that adapts to changing inputs and conditions.
Best Practices for Using Agent Nodes
To get the most out of the Agent Node, follow these best practices:
- Choose the right agent: Select an agent that's well-suited to the task at hand.
- Provide clear instructions: Make sure your instruction is clear, concise, and well-formed. This will help the agent produce high-quality output.
- Mention the right variable: Make sure you mention the appropriate variable in the instruction of the agent node. This will help the agent have the right content to produce the desired output.
By mastering the Agent Node, you can unlock the full potential of MindPal Multi-Agent Workflow and create complex, automated workflows that drive business value.