A workflow is a sequence of connected steps that process data. You build workflows visually using a drag-and-drop editor, then run them on demand or on a schedule. To get started, navigate to the Workflows page and click "Create New Workflow."

There are four ways to create a workflow:

• Blank: Start from an empty canvas. You choose a title, and the editor opens with no steps. Good when you know exactly what you want to build.
• From a Template: Pick a pre-built workflow from the template gallery. The editor opens with steps, edges, and field sets already configured. You can modify anything after creation.
• Import JSON: Paste a workflow JSON export to recreate a workflow from another account or environment.
• AI Generate: Describe what you want in plain language, and AI builds a workflow for you. You can review and edit the result before running it.

In addition to generating a workflow from scratch, you can use AI to modify an existing workflow. From the editor, describe the changes you want, like adding a new step, rearranging the flow, or updating field mappings, and AI will apply them for you. This is useful for making quick adjustments without manually reconfiguring steps.

Once your workflow is created, you land in the visual editor. The editor has three main areas:

• Step Palette (left sidebar): Lists all available step types organized by category. Drag a step onto the canvas to add it to your workflow.
• Canvas (center): The visual graph where your steps appear as nodes. Connect steps by dragging edges between them to define execution order.
• Step Configuration (right panel): Opens when you select a step. Here you configure the step's settings, define its input and output field sets, and set up field mappings from upstream steps.

1. Drag a step type from the palette onto the canvas. The step appears as a new node.
2. Click a step to open its configuration panel. Give it a title, configure its settings, and define input/output field sets.
3. Connect steps by dragging from one step's output handle to another step's input handle. This creates an edge that controls execution order.
4. Set up field mappings to pass data between steps. In a step's input field set, map each field to an output field from an upstream step.

Each step has a configuration panel with several sections:

• Title: A descriptive name for this step in the workflow.
• Settings: Step-type-specific options (e.g., which AI model to use, which Slack channel to post to).
• Input Field Set: The fields this step expects as input. You map these from upstream steps' outputs using field mappings.
• Output Field Set: The fields this step produces. Downstream steps can map from these outputs.
• Setup Errors: If a step has configuration issues (e.g., missing required settings, unconnected field mappings, or a missing integration), setup errors appear in the right panel. Resolve all setup errors before running the workflow.

Each step type has its own settings, required fields, and behavior. For detailed documentation on every step type, including what it does, how to configure it, and tips for common use cases, see the Step Types Guide.

Once you've added your steps, connected them with edges, and configured field mappings, your workflow is ready to run. See the Run a Workflow guide to learn how to execute your workflow and view results.