🇬🇧 EN

⌨️ Manual Input

Pauses the workflow and presents a data entry form to a specified user. The workflow resumes once the user submits the form. Use Manual Input when you need a human to supply data mid-workflow — for example, providing a correction, making a selection from options, or entering information that cannot be automated.

Category: Human / User Interaction · Type identifier: manual_input

Overview

The Manual Input node introduces a structured human data-entry step into an automated workflow. When execution reaches this node, the workflow enters Waiting status and a form task appears on the Tasks page for the assigned user. The user fills in the configured fields and clicks Submit. The submitted values are then available to all subsequent steps in the workflow context.

Unlike the Approval node (which only asks for a yes/no decision), Manual Input lets you collect arbitrary structured data — free text, numbers, selections from a dropdown, email addresses, and more — before the workflow continues.

Configuration

Field	Status	Description
Assignee	Required	Email address or user ID of the person to request input from. Supports `{{ variable }}` references.
Title	Required	Heading shown on the input form, e.g. `Review and confirm the extracted data`.
Description	Optional	Instructions or context for the user filling out the form.
Fields	Required	List of form fields to show. Each field has: Label — display name shown to the user. Key — variable name used to access the submitted value in outputs. Type — `text`, `textarea`, `number`, `email`, `select`, or `checkbox`. Options (for select) — comma-separated list of options. Required — whether the field must be filled before submitting.
Timeout	Optional	Maximum wait time in minutes. The workflow fails if the timeout expires without a submission. Leave blank to wait indefinitely.
Send Email	Optional	Toggle. Send an email notification to the assignee in addition to the in-app notification. Requires SMTP to be configured in Settings.

Output Data

All submitted field values are available as a single data object keyed by the Key you configured for each field. Individual fields can be accessed directly using dot notation.

Field	Type	Description
`data`	object	All submitted field values as key-value pairs, keyed by the field Key you configured.
`data.[key]`	any	Individual field value, e.g. `{{ input.output.data.approved_amount }}`.
`submitted_by`	string	Email or user ID of the person who submitted the form.
`submitted_at`	string	ISO 8601 timestamp of the submission.

Example Usage

A Schema Extractor node automatically extracts invoice data (vendor name, amount, date) from a scanned PDF.
A Manual Input node presents the extracted values to a finance team member, asking them to verify the amount. The form has a number field with Key approved_amount and a textarea field with Key notes.
A Counter node updates the running total using {{ input.output.data.approved_amount }}.
An Email Notification node sends a confirmation to the accounts team with the verified amount and the submitter's notes.

Tips & Notes

Choose field Keys carefully — they become the variable names you reference in downstream steps (e.g. a Key of approved_amount is accessed as {{ input.output.data.approved_amount }}). Use lowercase, underscore-separated names with no spaces.
Use a select field with predefined options when you want to constrain the user's choices — for example, a department selector or a priority level.
Set a Timeout for time-sensitive workflows to prevent runs from hanging indefinitely if the assignee is unavailable.
The workflow remains in Waiting status on the Tasks page until the user submits the form, making it visible to administrators monitoring active runs.
Use the Description field to provide context, instructions, or data from earlier steps (using variable references) so the user understands what they are being asked to verify or enter.

Related Nodes

Approval — pause for a simple approve/reject decision rather than structured data entry.
In-App Notification — alert a user that their input is needed.
Schema Extractor — automatically extract structured data before asking a human to verify it.