Webhook Agents allow you to send data from your workflow to an external system and use the response in your workflow.
The Webhook Agent lets you connect your workflows to any external API using HTTP requests. You can send or retrieve data dynamically using information from earlier steps in your automation.
This guide explains how to configure the Webhook Agent and how to use dynamic values pulled from your workflow state to populate your requests.
Here’s a breakdown of each field you’ll configure:
Use this to describe what the agent should do. This prompt helps guide the AI’s reasoning and ensures it maps the right values into your request body.
Enter the full URL of the endpoint you want to call. For example:https://api.example.com/customers
Select the HTTP method to use for the request. Supported methods include:
GET
POST
PUT
DELETE
If the endpoint requires authentication, you can paste your access token here. It will be automatically added to the request headers as a bearer token:
makefile
CopyEdit
Authorization: Bearer <your-token>
This is where you define the JSON structure to be sent with your request. You’ll define:
The request body should follow valid JSON formatting. Each key represents a field the API expects. Each value can either be:
"status": "active"
)"email": "Enter the customer’s email address from the previous step"
)This makes it possible for the agent to fill in request data using context from earlier in the workflow.
{
"email": "Enter the customer’s email address from the previous step",
"name": "Enter the full name of the customer",
"status": "active"
}
In this example:
email
and name
will be filled in by the agent based on earlier steps.status
will be sent exactly as written: "active"
.⚠️ Make sure the keys ("email"
, "name"
, "status"
, etc.) match what your target API expects. The agent will only update the values, not the keys.
Here are a few ways you can use the Webhook Agent:
The Webhook Agent is one of the most flexible tools in your Victoria AI toolkit. With the ability to dynamically populate request bodies and connect to any external system, it opens the door to powerful integrations without writing any code.
By combining natural language prompts with structured data, you can build smart, context-aware workflows that connect across your entire stack.
Let us know if you’d like help with your first integration — or check out our templates to get started faster.