Back to Blog

Introducing Aiflowly: no-code AI flow and agent builder

Airflowly is a no-code AI flow and agent builder. It runs locally on your browser and works with most AI models.

Posted by

Demonstration of Aiflowly workflow with a no-code text-to-text and text-to-image AI models.

Introduction

Airflowly is a no-code AI flow and agent builder. It runs locally on your browser and works with most AI models. You can use it to build AI flows and agents without writing any code. It's perfect for developers, data scientists, and AI enthusiasts.

Visual Flow Designer

Airflowly comes with a visual flow designer that lets you create AI flows and agents by connecting nodes. You can use it to build complex AI flows and agents without writing any code.


Demonstration of Aiflowly flow with visual flow designer.

Model Parameters

Each AI model has its own distinct set of parameters. You can configure these parameters to customize the behavior of the model.


Part of the model parameters are the input and output fields. You can specify the input and output fields of the model to match the input and output fields of the AI flow.


For example, if you are building an AI flow with a text-to-text model, input for the model can be a static or user provided text, or it can be the output of another model in the flow.


Similarly, the output of the model can linked to another model in the flow, such as a text-to-image model. In this case text-to-image model can use it as a prompt to generate an image.


Example Flow

Here is an example of a simple AI flow that uses a static-text-input, text-to-text, and text-to-image models.


Example of a simple AI flow with static-text-input, text-to-text, and text-to-image models.

Input

The input of the flow is a static text that is provided by the user. It is as simple as: generate a random topic.


Text-to-Text Model

The text-to-text model takes the input text as a prompt and generates an input prompt for the text-to-image model.


To instruct the text-to-text model to generate a prompt for the text-to-image, we use the system message:

generate an input prompt for text to image AI model based on provided input

Text-to-Image Models

To demonstrate the capabilities of the Aiflowly platform, we used two different text-to-image models to generate images based on the provided input:



Here are the results of these models:


SDXL Model
Example result of the text-to-text model.
Kandinsky 2.2 Model
Example result of the text-to-text model.

Notes

We did not use any custom parameters for the models in this example. The models were used with their default parameters.