Having a streamlined workflow can save you time and increase productivity. In this guide, we’ll walk you through the essential steps for creating a workflow that works for you.
There are 3 steps to creating a workflow:
- Build your workflow
- Test your workflow
- Activate your workflow.
Step 1: Build your workflow
Create a workflow
-
On the sidebar, click on the “WorkFlows” option.

-
Press on the “New Workflow” button located at the top bar.

-
After clicking “New Workflow”, choose one of those two options:

- Create- create a workflow manually.
- Import- import an existing workflow saved as a JSON file.
Choose a Trigger
- Follow the instructions above to create a new workflow.
-
Click “Add first step…” to choose a trigger that will start the process.

-
App Event - In the what’s trigger your flow activity block presented to you, select the software you want to use for your trigger. You can select the software by either clicking on one of the icons presented to you, or by filtering the requested software and selecting the requested option. You can also filter the trigger activities directly.

-
You can choose the trigger activity from other available options

- Manually - This trigger allows actions that are initiated directly by a user clicking a button, or by an API call or a Run Workflow step. You can define the data it receives, see Manually trigger inputs.
- Schedule - This trigger supports actions that run automatically based on time.
- Webhook - This trigger enables actions that are executed when an external system sends an HTTP request. For more information about Webhooks
- MCP Tool Trigger- Allows you to initiate an automated workflow by passing structured data into your system. It functions as the starting point (or entry node) for complex data integrations, letting you define exactly what information the workflow expects to receive. Learn more about the MCP Tool Trigger.
-
App Event - In the what’s trigger your flow activity block presented to you, select the software you want to use for your trigger. You can select the software by either clicking on one of the icons presented to you, or by filtering the requested software and selecting the requested option. You can also filter the trigger activities directly.
Manually trigger inputs
The Manually trigger appears as a Button step on the canvas. By default it has one input,text, and whatever the caller sends is placed under it, so later steps read ##1.text. If the workflow is started with structured data, for example an order with a customer and a list of lines, you can define that data on the step, so its fields are available to every later step.
-
Click the Button step, then click Define manually to add the fields one by one, or Generate from sample JSON to paste an example of the data and let Engini build the fields from it.

- Generate from sample JSON - Opens a window where you paste a sample JSON payload. Engini builds the fields from it. The sample values themselves are not saved, only the structure. The sample must be a JSON object that contains at least one field.
- Define manually - Starts from an empty list. Nothing is saved on the step until you add the first field.
-
Once the inputs are defined, they appear as a tree under Inputs.

- Regenerate - Build the fields again from a new sample JSON.
- Remove schema - Removes all the inputs immediately, with no confirmation, and returns the step to the single text value. Removing the last field in the tree leaves the editor empty and saves nothing.
- Add field - Add a field to the object or list it sits under. Each field has a Name, an optional Display name and Description, a Type, and a Required checkbox. Use simple names without spaces or dots, because the field name becomes part of the reference.
string,number,boolean,date,datetime,object,arrayandfile. Anarrayadds an Element Type select, which lists the same types exceptfile. Afilefield adds Multiple. The top level offers onlyobjectorarray, and inputs can be nested up to 3 levels. Required is not enforced: a call that omits a required field is accepted without an error. It only marks the field in the editor and in a Run Workflow step. -
Use the fields in the next steps like any other value. For example,
##1.customer.name, where1is the Button step number.
- The data is sent to the workflow in the body of the call, for example from the Engini API or from a Run Workflow step. Running the workflow from the designer sends no data, so steps that read an input fail with “Can’t resolve field”. Use the API with a non-empty body to test: an empty
{}body stores no trigger data and fails the same way as a designer run. - Engini keeps
##1.textworking only if the step still holds the text input or another step references##1.text. In that case runs store the body twice (at the root and under##1.text), which roughly doubles that workflow’s history storage. A workflow that never used the text value stores the body once. - Rerunning a run that was recorded before you defined the inputs replays its old body, so steps that read
##1.<field>fail on it. - The dynamic content picker no longer shows a
textnode.##1.textstill works where it is kept, but it cannot be picked from the list. - Declaring a field called
textreplaces the old##1.textvalue. Engini asks you to confirm only when the workflow still uses it, because steps that use##1.text.<something>will then fail. - The Button step no longer offers the Add Field picker; its fields section is headed Inputs.
Add your Actions
-
Click on the “+” button to add an action activity block.

-
Select the type of activity you want to use for your action.
You can review the different options here

- App Action- In this case, it refers to using one of our existing connectors. You can find additional explanations for each one under the Apps section.
- AI Agents - Includes AI-powered actions that allow you to analyze text, extract information, generate responses, classify data, summarize content, or make intelligent decisions within the workflow. For more information click AI Agents
- Core - Provides fundamental system-level actions such as sending HTTP requests, handling data operations, interacting with internal services, or performing essential backend tasks needed to support integrations. For more information click on one of the next links:
- Flow Controls - Includes logic and control mechanisms inside workflows, such as conditions (if/else), loops, delays, branching, and error handling. These actions help manage how and when different parts of the flow execute. For more information click Flow Control
- Transform & Format - Contains tools for data manipulation, including parsing, formatting, mapping, converting data types, restructuring JSON, and preparing data before sending it to another system. For more information click Transform & Format
- Add Another Trigger - Allows you to add an additional trigger to the same flow, enabling the workflow to start from multiple different events or entry points. For more information click Triggers
Name your Workflow

- Click the “New Workflow” phrase located in the top left of the page.
- Remove the current name and enter the new name for the workflow. It is recommended to give a meaningful name to the workflow.
- Save the workflow.
Step 2: Test your workflow
Test your workflow to ensure that all components are functioning correctly.
Step 3: Activate your workflow
Your workflow is almost ready to go! Engini’s workflows are enabled by default. Make sure the on/off switch is activated.