Enhance your Agent’s capabilities by adding purpose-built Tools. Below is a quick start directory of all available tools, grouped by focus, with direct links to their detailed guides.


Tool Implementation

Tools can be added to an agent to extend its functionality. Each tool has a type and a config object that defines its behavior.

HTTP Tool

The HTTP tool allows the agent to make HTTP requests to external APIs.

Configuration:

  • url: The URL of the API endpoint.
  • method: The HTTP method to use (GET, POST, PUT, DELETE, PATCH).
  • headers: An array of key-value pairs for the request headers.
  • body: An array of key-value pairs for the request body.
  • rawBody: A raw JSON string for the request body.
  • queryParameters: An array of key-value pairs for the URL query parameters.
  • pathVariables: An array of key-value pairs for the URL path variables.

Datastore Tool

The Datastore tool allows the agent to query a datastore for information.

Configuration:

  • datastoreId: The ID of the datastore to query.

Mark as Resolved Tool

This tool marks the conversation as resolved. It has no configuration.

Request Human Tool

This tool requests a human to take over the conversation. It has no configuration.

Delayed Responses Tool

This tool adds a delay before the agent responds, to make the conversation feel more natural.

Configuration:

  • delay: The delay in seconds.

Follow-up Messages Tool

This tool sends follow-up messages to the user if they stop responding.

Configuration:

  • max_sends: The maximum number of follow-up messages to send.
  • interval_hours: The interval in hours between follow-up messages.
  • messages: The messages to send.

Best Practices

  • Keep tool descriptions short and action-oriented so the LLM knows exactly when to use them.
  • Mark variables as “Provided by User” when the value should be inferred from conversation context (HTTP Tool).
  • In Raw Mode, define types, items and properties explicitly to avoid malformed payloads.
  • Always preview ParsingEditor output to ensure only the necessary fields are exposed to the agent.
  • For follow-ups, monitor impact and adjust strategy to balance engagement and credit usage.