Multi-Step Pipelines (sequence)

This example demonstrates how to create a sequential pipeline that orchestrates multiple workflows to accomplish a complete automation goal. We’ll build a blog creation pipeline that researches a topic, writes content, and publishes it.

Note: This example uses create_mock tools for demonstration purposes to keep the example simple and focused on core workflow concepts. In real-world use cases, you would replace these mock tools with actual tools that perform the specific processing operations your automation requires (such as API calls, data transformations, file operations, or LLM interactions).

Pipeline Configuration

The root pipeline defines the sequence of workflows that execute one after another:


pipelines:
  - name: sequence_example
    title: "Example 1: Pipeline Sequence"
    type: root
    workspace: examples
    sequence:
      - workflow: research_topic      # Step 1: Research
      - workflow: write_content       # Step 2: Write content
      - workflow: publish_content     # Step 3: Publish

Key Concepts:

Sequential Execution: Each workflow completes before the next begins
Root Pipeline: Marked as type: root so it can be executed from Loopstack Studio
Workspace Assignment: All workflows inherit the examples workspace

Workflow 1: Research Topic

The first workflow uses mock data to simulate topic research:


workflows:
  - name: research_topic
    title: 'Research'
    type: stateMachine
    transitions:
      - name: research_topic
        from: start
        to: research_finished
        call:
          - tool: create_mock   # Mock tool for demonstration
            arguments:
              input: 'AI trends 2025'
              output: 'Key AI trends include: agentic AI systems, reasoning models, and AI-powered robotics.'
            exportContext: RESEARCH_RESULT       # Available to subsequent workflows
 
      - name: show_results
        from: research_finished
        to: end
        call:
          - tool: create_chat_message
            arguments:
              role: 'assistant'
              content: ${ context.variables.RESEARCH_RESULT }  # Access from context

Key Concepts:

Mock Tools: create_mock simulates tool behavior for testing and examples
exportContext: Makes data available to subsequent workflows in the pipeline sequence
Context Access: Use ${ context.variables.VARIABLE_NAME } to access data from context object
Chat Messages: create_chat_message displays results in the Studio interface

Workflow 2: Write Content

The second workflow accesses data from the previous workflow and creates content:


workflows:
  - name: write_content
    title: "Write Content"
    type: stateMachine
    transitions:
      - name: write_content
        from: start
        to: content_created
        call:
          - tool: create_mock
            arguments:
              input: ${ context.variables.RESEARCH_RESULT }  # Access from pipeline context
              output: |
                # Top AI Trends in 2025
                Here are the key trends defining this year...
            exportContext: CREATED_CONTENT
 
      - name: show_results
        from: content_created
        to: end
        call:
          - tool: create_chat_message
            arguments:
              role: 'assistant'
              content: ${ context.variables.CREATED_CONTENT }  # Access from context

Key Concepts:

Context Access: Use ${ context.variables.RESEARCH_RESULT } to access data from other workflows
Data Flow: Research results flow from the first workflow to inform content creation
Multi-line Output: Use | for multi-line YAML strings in mock tool outputs

Workflow 3: Publish Content

The final workflow publishes the created content:


workflows:
  - name: publish_content
    title: "Publish"
    type: stateMachine
    transitions:
      - name: publish_content
        from: start
        to: published
        call:
          - tool: create_mock
            arguments:
              input: ${ context.variables.CREATED_CONTENT }  # Access content from previous workflow
              output: 'Blog post "Top AI Trends in 2025" published successfully to blog.example.com'
            as: SUCCESS_MESSAGE      # Only available within this workflow
 
      - name: show_results
        from: published
        to: end
        call:
          - tool: create_chat_message
            arguments:
              role: 'assistant'
              content: ${ SUCCESS_MESSAGE }       # Access local variable

Key Concepts:

as: Use ${ SUCCESS_MESSAGE } to transfer data within the same workflow
Local Variables: Variables exported with as are only accessible within the current workflow

Data Flow Summary

This example demonstrates the complete data flow pattern in sequential pipelines:

Data Export Patterns:

exportContext: RESEARCH_RESULT → Access as ${ context.variables.RESEARCH_RESULT }
as: SUCCESS_MESSAGE → Access as ${ SUCCESS_MESSAGE } (within same workflow only)

Running This Example

This example is already available in Loopstack Studio:

Navigate to the Studio interface
Switch to the examples workspace
Select “Example 1: Pipeline Sequence” from the available automations
Click “Start” to execute the complete sequence
Watch each workflow execute in order, displaying results in the chat

Configuration Location: You can also view the complete configuration file at src/config/examples/basic/sequence-example.yaml

Complete Example:


include:
  - core/tools/create-mock.yaml
  - core/tools/create-chat-message.yaml
 
pipelines:
  - name: sequence_example
    title: "Example 1: Pipeline Sequence"
    type: root
    workspace: examples
    sequence:
      - workflow: research_topic      # Step 1: Research
      - workflow: write_content       # Step 2: Write content
      - workflow: publish_content     # Step 3: Publish
 
workflows:
  - name: research_topic
    title: 'Research'
    type: stateMachine
    transitions:
      - name: research_topic
        from: start
        to: research_finished
        call:
          - tool: create_mock   # A mock tool call returning the research result
            arguments:
              input: 'AI trends 2025'
              output: 'Key AI trends include: agentic AI systems, reasoning models, and AI-powered robotics.'
            exportContext: RESEARCH_RESULT
      - name: show_results
        from: research_finished
        to: end
        call:
          - tool: create_chat_message    # Display the results in the chat
            arguments:
              role: 'assistant'
              content: ${ context.variables.RESEARCH_RESULT }
 
  - name: write_content
    title: "Write Content"
    type: stateMachine
    transitions:
      - name: write_content
        from: start
        to: content_created
        call:
          - tool: create_mock   # A mock tool call returning the created content
            arguments:
              input: ${ context.variables.RESEARCH_RESULT }
              output: |
                # Top AI Trends in 2025 
                Here are the key trends defining this year...
            exportContext: CREATED_CONTENT
      - name: show_results
        from: content_created
        to: end
        call:
          - tool: create_chat_message    # Display the content in the chat
            arguments:
              role: 'assistant'
              content: ${ context.variables.CREATED_CONTENT }
 
  - name: publish_content
    title: "Publish"
    type: stateMachine
    transitions:
      - name: publish_content
        from: start
        to: published
        call:
          - tool: create_mock   # A mock tool call returning a success message
            arguments:
              input: ${ context.variables.CREATED_CONTENT }
              output: 'Blog post "Top AI Trends in 2025" published successfully to blog.example.com'
            as: SUCCESS_MESSAGE
      - name: show_results
        from: published
        to: end
        call:
          - tool: create_chat_message    # Display the success message in the chat
            arguments:
              role: 'assistant'
              content: ${ SUCCESS_MESSAGE }