Everything AIInstall

Summary

Generate and execute custom code based on natural language instructions to process data from multiple input ports and route results to multiple output ports dynamically.

Inputs

• Number of Inputs (required) — Select the number of input ports (1-10). This determines how many distinct input data streams the node will accept.
• Number of Outputs (required) — Select the number of output ports (1-10). This sets how many separate outputs the node will produce.
• Your Requirement (required) — Describe your data processing logic in natural language using numbers for inputs (1,2,3,…) and uppercase letters for outputs (A,B,C,…). The node uses this to generate the processing code.
• Data Complexity Level — Choose how much actual data (vs. just structure) is provided to the AI when generating code, ranging from structure only to full data (higher levels increase token usage).
• Model Name or ID (required) — Select or specify the large language model to use for generating processing code.
• Custom Model Name (required) — Specify the custom model name if "Custom" is selected as the model.
• Advanced Settings — Options including a custom prompt template, security check enabling, package configuration, and Playwright browser automation settings for remote code execution.

Output shape

An array where each element corresponds to an output port containing an array of data items; keys are uppercase letters (A, B, C, ...) matching the output ports, and each key maps to a list of items to emit on that output.

The node dynamically generates JavaScript code from the natural language instruction to process inputs and define outputs. This code must return an object mapping output port letters to arrays of data. The node enforces security checks to block dangerous operations by default. If Playwright browser automation is enabled, execution may occur remotely and can maintain browser context. Errors in code generation or execution will produce descriptive node errors.