Bink AI Agent

Initialize Bink AI Agent with plugins
Join our community

Overview

This node initializes and runs a Bink AI Agent that can process user input using different agent types, such as a Tools Agent or a Plan and Execute Agent. It supports structured tool usage, memory integration, and optional output parsing to format the response. The node is useful for automating complex tasks that require AI-driven tool selection, execution, and stepwise problem solving, such as bridging tokens across multiple blockchain networks or executing multi-step workflows.

Use Case Examples

  1. Using the Tools Agent to query token information and perform blockchain bridging operations with precise tool calls.
  2. Using the Plan and Execute Agent to solve complex tasks by planning and executing steps sequentially, with customizable human message templates.

Properties

Name Meaning
notice_tip Displays a tip with links to tutorials and examples for users of the Tools Agent.
hasOutputParser Boolean flag to require a specific output format by connecting an output parser node.
notice Notice shown when an output parser is connected, explaining how to specify the output format.
promptType Source for the user input message, either from a connected Chat Trigger node or defined manually.
text The user's input text to be processed by the agent.
agent Selects the type of AI agent to use for processing the input.
options Additional configuration options for the selected agent, such as system messages, max iterations, and whether to return intermediate steps or passthrough binary images.

Output

JSON

  • json
    • output - The main output generated by the AI agent after processing the input.
    • intermediateSteps - Optional intermediate steps taken by the agent if enabled.
    • error - Error message if the node fails to process an item and continueOnFail is enabled.

Dependencies

  • Binkai API Credentials for blockchain network access and wallet management

Troubleshooting

  • If the 'text' input parameter is empty, the node throws an error indicating the missing input; ensure the input is correctly provided.
  • Errors loading memory variables may occur; these are logged but do not stop execution.
  • If the node fails to process an item and continueOnFail is disabled, the workflow will stop with an error; enable continueOnFail to handle errors gracefully.

Links

Discussion