Bro AI Agent

Gera um plano de ação e o executa. Pode usar ferramentas externas.
Join our community

Overview

This node, named 'Bro AI Agent', is designed to generate and execute an action plan using an AI agent that can utilize external tools. It is typically used in scenarios where automated decision-making or task execution is needed, leveraging AI language models and optionally memory, tools, and output parsers. For example, it can be connected to a compatible language model node to process user prompts and perform complex workflows involving multiple steps and external tool integrations.

Use Case Examples

  1. Automating customer support responses by generating action plans based on user queries.
  2. Executing multi-step data processing tasks where the agent decides the sequence of operations.
  3. Integrating with external APIs or services to perform tasks based on AI-generated instructions.

Properties

Name Meaning
Agent Specifies the AI agent type to use, defaulting to 'toolsAgent'.
Fonte para Prompt (Mensagem do Usuário) Defines how the user prompt is provided, either from a connected chat trigger node or directly defined in this node.
Prompt (Mensagem do Usuário) The user message text sent to the AI model, either dynamically from previous node data or statically defined here.
Requerer Formato de Saída Específico Indicates whether a specific output format is required by connecting an output parser.
Opções Additional options for the agent's behavior, including system message, max iterations, returning intermediate steps, and passing binary images automatically.

Output

JSON

  • json
    • error - Error message if the agent execution fails.
    • details - Error stack trace details if available.
    • result - The final output or response generated by the agent after executing the action plan.
    • intermediateSteps - Optional intermediate steps taken by the agent if enabled in options.

Dependencies

  • Requires connection to a compatible AI language model node for prompt processing.
  • Optionally uses memory, tools, and output parsers as connected nodes.

Troubleshooting

  • Common errors include failure to execute the agent due to missing or incompatible AI model connections.
  • Error messages provide details and stack traces to help diagnose issues.
  • Ensure the 'agent' parameter is correctly set and that the connected language model node is properly configured.
  • If 'continueOnFail' is enabled, the node returns error details in JSON instead of failing the workflow.

Links

Discussion