CodeAct Agent Node

Execute Python code dynamically to complete complex analytical and computational tasks

Overview

The CodeAct Agent Node provides an AI agent that can write and execute Python code to solve complex problems. This node enables:

• Data analysis and visualization
• File processing and manipulation
• Mathematical computations and modeling
• API interactions and data transformations
• Multi-step programming tasks with iterative refinement

Configuration Parameters

To set up the CodeAct Agent Node, configure the agent’s behavior and select the tools it can use alongside code execution.

Node Configuration

• Query:

  Define the task for the agent to complete using Python code.

  Analyze this dataset and create visualizations showing key trends

  Process these CSV files and generate a summary report

  Calculate statistical metrics and export the results

• Tools

  Select additional workflow or node tools that the CodeAct Agent can use alongside code execution. You can combine code execution with other integrations and capabilities.

  The agent can execute Python code and use the selected tools together to complete complex tasks.

Advanced Settings

• System Prompt: Customize the agent’s coding behavior and guidelines.

• Model: Select the language model for the agent (must support function calling).

• Max Steps: Set the maximum number of steps the agent can take (1-50).

• Enable Model Fallbacks: Allow automatic fallback to alternative models if the primary model fails.

• Inject Access Token Function: Inject the get_access_token function into the sandbox for retrieving integration credentials.

Expected Inputs and Outputs

• Inputs:

  • input: Text input that can be referenced in the query using format strings
  • files: Files to use in the CodeAct agent sandbox

• Outputs:

  • output: The final result from the agent’s code execution
  • conversation: Detailed log of the agent’s reasoning and code execution steps
  • files: Files generated during code execution in the sandbox

Use Case Examples

1. Data Analysis:

   Use the agent to analyze datasets, calculate statistics, and generate insights automatically using pandas, numpy, and other data science libraries.

2. Report Generation:

   Configure the agent to process input data, perform calculations, and generate formatted reports with charts and visualizations.

3. File Processing Automation:

   Let the agent read, transform, and process files using Python code, handling various file formats and data transformations.

Error Handling and Troubleshooting

• Model Compatibility:

  The selected model must support function calling. Switch to a compatible model if you encounter errors.

• Code Execution Errors:

  The agent will attempt to debug and fix code errors automatically. Review the conversation output to see the agent’s problem-solving process.

• Sandbox Limitations:

  Code execution happens in a sandboxed environment with limited access to external resources. Network requests and file system access may be restricted.

If you encounter any issues not covered in this documentation, please reach out to our support team for assistance.

Relevant Nodes

Agent Node

General-purpose agent for complex tasks

LLM Node

Direct access to language models for text generation

Document Reader Node

Extract text content from various document formats