Introduction to Digital Employee (DE) Pipeline

Digital Employee Pipeline (DE Pipeline) is a framework for building AI agent workflows. It allows you to create processing pipelines by combining reusable components.

Why a Pipeline Is Introduced

1. Efficiency: Fewer LLM calls — reduced operational cost

2. Maintainability: Modular step design — independent changes without cascading effects

3. Reproducibility: Deterministic execution — same inputs produce same outputs

4. Responsiveness: Fewer LLM calls — reduced end-to-end latency

5. Testability: Isolated step boundaries — independent testing of each step

Start Here

The fastest path is a minimal Digital Employee:

from digital_employee_core import (
    DigitalEmployeeAgentBuilder,
    DigitalEmployeeBuilder,
    DigitalEmployeePipelineBuilder,
)

# Build and run
digital_employee = (
    DigitalEmployeeBuilder()
    .set_name("My Assistant")
    .set_email("[e-mail]")
    .set_job(
        title="Assistant",
        description="A helpful assistant",
        instruction="You are a helpful assistant.",
    )
    .set_pipeline(
        DigitalEmployeePipelineBuilder().add_step(
            DigitalEmployeeAgentBuilder()
        )
    )
    .build()
)

result = asyncio.run(digital_employee.invoke(message="Hello!")

Recommended Reading Path

• Getting started: install, configure, and quick start.

• Architecture: understand components, steps, and pipelines.

• Guides: build multi-step pipelines, extend with custom components.

• Advanced examples: explore real-world patterns and use cases.

Documentation Map

Use these sections in order when exploring DE Pipeline:

• Getting Started — Install, configure, and run your first Digital Employee.

• Architecture — Understand the architecture: components, steps, pipelines, and state management.

• Guides — Deep dives on multi-step pipelines, extending with custom components, tools, and MCPs.

• Advanced Examples — Real-world patterns including preprocessing, postprocessing, and complex workflows.

Pipeline Patterns at a Glance

Choose the pattern that matches your requirements:

Pattern	Description	When to use it	Documentation
Single Agent	Simple Digital Employee with one processing step	Basic conversational agents	Getting Started
Multi-Step Pipeline	Chain multiple processing steps together	Add preprocessing/postprocessing, validation, transformation	Multi-Step Pipelines
Extended Pipeline	Custom components, tools, and MCPs	Advanced workflows with external integrations	Extend

Architecture

DE Pipeline is built on three core concepts:

Components

Components are the building blocks of pipelines. Each component processes input and produces output, implementing specific functionality like validation, transformation, or formatting.

Steps

Steps wrap components with input/output mappings, defining how data flows through the pipeline and how state is managed between steps.

Pipelines

Pipelines chain multiple steps together to create a complete processing workflow. Data flows through each step sequentially, with the output of one step becoming the input for the next.

Input → Step 1 → Step 2 → Step 3 → Output

Key Concepts

Flexibility - Mix and match components as needed, combine any number of steps in any order.

Simplicity - Each step is self-contained and testable, making pipelines easy to understand and maintain.

Composability - Build complex workflows from simple components and reuse them across different pipelines.

Start Building

Ready to build your first Digital Employee? Follow this path:

1. Install & configure — Set up dependencies with Getting Started.

2. Run the quick start — Create your first Digital Employee and understand the basics.

3. Explore multi-step pipelines — Learn how to chain steps together with Multi-Step Pipelines.

4. Extend your pipeline — Add custom components, tools, and MCPs with Extend.

5. Explore advanced examples — See real-world patterns in Advanced Examples.

Digital Employee Pipeline gives you a flexible, composable framework for building sophisticated AI agent workflows.