Webhook Triggers
Audience: Administrators, Solution Architects, and Developers
Purpose: Documents the Webhook trigger in Kizen: how it works, how data is extracted from inbound requests, and the behavioral details that affect how webhook-triggered Agentic Workflows are designed and debugged.
Overview
The Webhook trigger initiates an Agentic Workflow via an inbound HTTP request from an external system. Each webhook trigger generates a unique URL that external tools and services can call directly to start an Agentic Workflow execution.
Webhook URL
Each Webhook trigger is assigned a unique URL based on the webhook name configured at setup.
The URL follows this format:
The webhook name is user-defined and becomes part of the URL. Choose a name that clearly identifies the Agentic Workflow and integration context and it cannot be changed after the trigger is saved without breaking existing integrations.
POST and GET Requests
The Webhook trigger supports both POST and GET requests. The selected method determines what data extraction options are available.
POST
POST requests support a request body payload. When POST is selected, the trigger configuration exposes an example Payload field, a Content-Type selector, and data extraction options for both the request body and URL query string.
Content-Types
When using POST, the Content-Type of the inbound request can be set to match the format sent by the external system. Supported content types include:
application/jsonapplication/javascriptapplication/xmlapplication/xhtml+xmlapplication/x-www-form-urlencodedtext/plaintext/htmltext/xmltext/csv
GET
GET requests do not include a request body. When GET is selected, only URL query string extraction is available.
Record Identifier Requirements
Inbound webhook requests must include a Record identifier so Kizen can associate the request with the correct Record and initiate the Agentic Workflow in the appropriate context. This is only applicable for Record-based Agentic Workflows, not global ones.
Data Extraction
The Data Extraction section controls what information is pulled from the inbound request and made available as variable sources within the Agentic Workflow. Available options differ by request method.
JSON Path (POST only). Individual values can be extracted from the request body using JSONPath expressions. Each extraction row requires a JSON Path expression, a Name that becomes the variable identifier downstream, and optionally an Example Payload to preview the extracted value at configuration time. Rows can be added manually with + Add Value or generated automatically from the Example Payload using Generate Values Automatically. This is only applicable if the content-type is JSON.
Extract Full Body Content (POST only). When enabled, the full request body is extracted and made available as a variable source. Use this when the entire payload is needed downstream rather than specific fields.
Extract URL Query String (POST and GET). When enabled, values from the URL query string are extracted and made available as variable sources. This option is available for both POST and GET requests.
All extracted values flow into the Agentic Workflow's variable system and can be referenced by any subsequent step that supports variable inputs.
Webhook Triggers Cannot Fail Directly
The Webhook trigger itself does not fail. Kizen always accepts the inbound request regardless of what the payload contains.
If a required Agentic Workflow variable depends on a value not present in the payload, the failure surfaces at the variable evaluation stage and not at the trigger. When debugging a webhook-triggered Agentic Workflow that is not behaving as expected, check variable initialization first.
What's Next?
Now that you understand how each trigger type works, continue to Agentic Workflow Code Steps to learn learn how to run custom Python scripts within an Agentic Workflow to handle complex logic, external API calls, and data transformations that go beyond built-in actions.
Last updated
Was this helpful?