# Agent Node

### Overview

The Agent Node empowers your flow with decision-making capabilities by creating an AI agent that can use multiple tools to accomplish tasks. It combines a large language model with a set of tools, enabling complex reasoning and multi-step task completion based on the input and available tools.

Usage cost: 2 credits

### Configuration

#### Settings

1. **Model Selection**
   * LLM Model\*: Select the AI model for the agent (the selected model must support tools calling)
   * System Prompt: Initial instructions defining the agent's behavior and context
   * Agent Prompt\*: The main instruction or task for the agent
   * Past Message History: Optional chat history for context
2. **Tools Configuration**
   * Tools\*: Select one or more tools for the agent to use
   * Tools must be provided by previous nodes in the flow
3. **Output format**

   * **Output as JSON** (Toggle):
     * **Off (Default):** The node outputs the model's response as a plain string.
     * **On:** Instructs the model to format its response as JSON and attempts to parse the output string. If parsing succeeds, the `response` output port will contain a JSON object/array. If it fails, `response` will contain the original string.

   *Note:* JSON structure and parsing success depend heavily on the model's ability to follow instructions. Clearly prompting for JSON format is recommended when this is enabled.

#### Output Ports

* `response` (string): The agent's complete response after using tools and reasoning

### Best Practices

1. **Tool Selection**
   * Only provide tools that are relevant to the task
   * Order tools from most to least commonly used
   * Ensure tools have clear, descriptive names
   * Limit the number of tools to prevent confusion
2. **Prompt Engineering**
   * Be specific about what tools should be used
   * Clearly define the expected outcome
3. **Performance Optimization**
   * Keep tool descriptions concise
   * Use appropriate model for complexity
   * Balance between instruction detail and flexibility

### Common Issues

* Tool execution timeouts
* Incorrect tool selection by agent
* Incomplete or unclear tool responses
* Model context length limitations