Custom mid-call tools allow your AI agent to interact with your external systems in real time, during a call. Inventory checks, customer data, real-time information—everything is possible without complex development.

Operating principle

You configure an API endpoint with its parameters, and the AI ​​agent automatically determines when and how to use it based on the flow of the conversation. No code is required on the platform side.

Configuring a tool

# Example: Order status check Name: check_order Endpoint: https://api.yourstore.com/orders/{order_number}/status Parameters: - order_number (string): "Order reference provided by the customer (format: CMD-XXXXX)" # The agent automatically replaces {order_number} # with the value collected during the conversation

Test your tool

Before deploying, use the "Test the tool" with dummy data. You'll see the response code and the returned content—ideal for validating that your endpoint is working correctly.

💡

For more complex business logic, combine your custom tools with the no-code automation platform to create complete workflows.