Workflow

Build AI workflows using the visual Flow Editor with nodes, tools, and integrations.

Create powerful AI workflows by connecting nodes in the visual Flow Editor. This guide covers node types, connections, and how to enable chat features.

  1. Open the sidebar menu
  2. Click Workflows | Agents to expand
  3. Select Workflows

Create a New Workflow

  1. Click New Flow
  2. Enter Name and Description
  3. Click Create

You'll be taken to the Flow Editor.

Flow Editor

Key Areas

AreaPurpose
CanvasDrag and connect nodes
Node PaletteAvailable node types
Properties PanelConfigure selected node
Action MenuImport/Export JSON
PublishDeploy and embed options

Node Types

Chat Input

Entry point for conversational workflows. Required for Chat Playground.

PropertyDescription
Variable NameInput variable (default: input)

Chat Output

Returns responses to users. Required for Chat Playground.

PropertyDescription
Output VariableResponse to display

Agent

AI-powered processing node with model and tools.

PropertyDescription
ModelSelect configured model provider
System MessageInstructions for the agent
Store StatePersist conversation context
ToolsGateway MCP Client, Knowledge Base, etc.

Important: To use models from Configurable Components, ensure both Stream and Configure toggles are enabled in the model configuration.

Gateway MCP Client

Connects external tools from MCP Gateway.

PropertyDescription
ServersSelect from My Servers (Gsuite, Hubspot, etc.)

Knowledge Base

Adds RAG capabilities with your documents.

PropertyDescription
Knowledge BaseSelect configured KB

Connecting Nodes

  1. Click output port (right side of node)
  2. Drag to input port (left side of target node)
  3. Release to create connection

Basic Pattern

[Chat Input] → [Agent] → [Chat Output]

With Tools

[Chat Input] → [Agent + Tools] → [Chat Output]
    [Gateway MCP Client] + [Knowledge Base]

Chat Playground

Important: To enable Chat Playground, your workflow MUST have both Chat Input and Chat Output nodes connected.

  1. Add Chat Input node
  2. Add Chat Output node
  3. Connect through your Agent
  4. Save the workflow
  5. Click Playground to test

Import/Export JSON

Export Workflow

  1. Open Action menu
  2. Click Export
  3. Download JSON file

Import Workflow

  1. Open Action menu
  2. Click Import
  3. Upload JSON or paste content
  4. Review and save

Publish & Embed

Deploy your workflow as a chat widget:

  1. Save your workflow
  2. Click Publish
  3. Select Embed option
  4. Copy the embed code
  5. Add to your website
<iframe src="https://lasius.io/embed/{workflow-id}" />

Examples


Next: Knowledge Base