Manual Trigger

Starts a workflow on demand — from the canvas editor, the CLI, or when called as a subworkflow by another workflow.

When to Use

Use Manual Trigger when:

You are building a workflow that you want to call as a subworkflow from another workflow
You want to run a workflow manually from the canvas during development
You want to trigger execution on demand via tensorify run without needing an HTTP request

If you need a workflow that responds to external HTTP events, use Webhook or API Endpoint instead.

Handle	Type	Description
`payload`	object	The invocation envelope

The payload object shape:

Field	Type	Description
`body`	any	The data passed to this trigger (from a subworkflow invocation or test payload)
`headers`	object	Headers from the invocation
`invocation`	object	Metadata about how the workflow was invoked
`source`	string	`"manual"` or `"subworkflow"`

Setting	Type	Default	Description
`allowWorkflowInvocation`	boolean	`true`	Allow other workflows to call this as a subworkflow. Disable to restrict to manual runs only.
`mockPayload`	JSON	—	Test data injected when running this workflow from the canvas or via `tensorify watch`.

A basic subworkflow that receives data from its parent and transforms it:

Manual Trigger → Transform → Return

The parent workflow passes data to Manual Trigger via the Subworkflow node. The body field on payload contains that data.

In a downstream node, reference the trigger payload as:

{{ manual.body }}
{{ manual.body.someField }}

If allowWorkflowInvocation is disabled, calling this workflow from a Subworkflow node will fail at runtime.
The mockPayload is only used during canvas testing (tensorify watch). In production, the real invocation payload is used.