​
Introduction

Creating efficient workflows is crucial for handling complex tasks like data management, process automation, or running evaluations. A flexible, user-friendly system can save time and ensure better results.

Flows in Athina AI provide a low-code framework for building powerful pipelines that seamlessly connect tasks. By chaining together steps (called blocks), you can design workflows to process inputs, retrieve data, execute custom code, and more. Flows enable users to quickly prototype ideas, build reusable pipelines, and collaborate effectively.

This guide will introduce you to Flows, explain their core concepts, and provide a step-by-step walkthrough for creating and sharing them in Athina AI.

Let’s begin by understanding what Flows are.

​
What Are Flows?

Flows are a powerful way to create and manage complex pipelines by chaining together different steps, known as blocks, in a logical sequence. Each block performs a specific task, such as generating a response, running code, or retrieving data. These blocks work together to process inputs and produce outputs seamlessly.

Think of Flows as low-code notebooks for LLM engineering. Flows offer several key capabilities:

  • Build Multi-Step Workflows: Chain tasks like generating responses, retrieving data, and analyzing results efficiently.
  • Use Variables: Dynamically pass data between blocks to enable flexible and adaptable workflows.
  • Integrate with External Tools: Incorporate API calls or code execution blocks for advanced functionality and external integrations.
  • Notebook Interface: Designed as an easy-to-use notebook interface, Flows allow you to add headers, text, and comments, making them more readable and shareable.

With these features, Flows provides a versatile platform for designing reusable workflows tailored to a wide range of use cases, from prompt optimization to data analysis and more.

Now, let’s see the step-by-step implementation of a research paper summarizer Flow.

​
Implement Flows in Athina AI

​
Step 1: Add the Search Block

1

To start, open the Flow Builder in Athina AI and type a forward slash (/) to bring up the block menu. From this menu, select the Search Block.

​
Step 2: Enter Your Input Query

1

Once the Search Block is added, enter your query into the Input section. You can use this input dynamically in all your blocks like this {{query}} to retrieve results related to your input.

​
Step 3: Configure the Search Block

1

Next, configure the Search Block by adjusting the settings in the right-hand sidebar and clicking on Run Block.

​
Step 4: Add the Run Prompt Block

1

Once the Search Block is complete, add and run a Run Prompt Block to your Flow. This block will use an LLM to summarize or process the content retrieved by the Search Block, as shown below.

To get started quickly, you can explore the pre-built Flow Templates available in the library and customize them to your needs.

​
Step 5: Share Your Flow

1

Click on the Share button in the top-right corner of the Flow Builder. Adjust the sharing settings as needed to share with your team or publish it as a template in the Flow Templates Library, as shown below.

Following all of the above steps, you can create and share Flows with your team.

Implementing Flows in Athina AI allows you to design dynamic, automated workflows that simplify complex tasks. By using blocks like Search, Run Prompt, custom Code Execution, and more, you can create powerful pipelines tailored to your specific needs. Sharing your Flows fosters collaboration and reusability, making it easier to work within your team or contribute to the wider Athina AI community.

Compare Multiple Models