How to create a no-code AI app?

Learn how to integrate pre-trained models in efficiently.


Ever wondered, can integrate Deep Learning models in its applications, as Twitter and Alibaba do? If so, you are in the right place. is a no-code/low code platform, which integrates another such no code low code tool called “peltarion” to add deep learning models to its applications. Let’s dive into peltarion first.

Intro to peltarion

Peltarion is an AI platform which allows users to build their own learning models from scratch or they can use the pre-built ones and can customize them according to their needs. The best part is that you don’t need to code.

Using peltarion is pretty easy. It consists of the following steps

  • Sign Up
  • Create Project
  • Data Import
  • Inspect and build
  • Train and Evaluate
  • Deployment

Sign Up

Like any other online platform, you first need to sign up to peltarion and start your free trial of 4 weeks. You can subscribe later.

Create Project

After signing up, you are redirected to here.

You can directly create a project from here by selecting one of the algorithms or you can create your own by clicking on “start custom project”.

Once you create a project you now need data for it.

Data Import

To train a model we need data. There are two ways you can get that


You can get your data directly from peltarion too, if you created a project using the first method mentioned above peltarion provides you with a dataset already imported.


If you want to upload your own dataset that you created or downloaded from sources like “kaggle”, you can import those into your project too.
Here is an overall view of all the datasets you have in your account.

Inspect and Build

Once you have imported your data, it’s time to do some modelling to do that simply click on “use in the new experiment”

You name your experiment and decide what dataset to use.

what would be the Input/Output for the model.

Finally, you select the problem type from a predefined list provided by the platform. Which in our case is Single-label text classification.

Once you have created the experiment you can customize the build or leave it as is since peltarion shows you the best options for the built model.

Congratulations, you now have a deep learning model read for training.

Train and Evaluate

Time to Train our model. After you are done satisfied with the model setting click on the button “Run” on the top right corner

The training can take some time depending on your internet connection, dataset size etc.

You can update your training view by toggling among parameters like error, accuracy, learning rate etc

Once you are done with the first training experiment, the platform allows you to do cross-entropy to get the best results.


You can now select the best experiment for your case and deploy that. Using the button “create deployment”.

Keep this page open when you want to integrate this experiment into

Use case

Let’s have a look at the use case to understand the process. Let’s say some person X has built an app to sell his products online to keep track of what people thought about them, but Mr X doesn’t have much time on his hands to read all the reviews word by word. He wants the app to categorize the reviews into the following sentiments.

  • Positive
  • Negative
  • Neutral

Looking at the use case and having some knowledge about the deep learning algorithms, we can safely say that we need to integrate “sentiment analysis” into Mr X’s application.
Let’s refresh what sentiment analysis is.

Intro to Sentiment analysis

The dictionary definition is that it is a process of computer identification and classification of ideas expressed in the form of a text, especially to determine the author’s attitude towards a particular topic, product, etc. positive, negative, or neutral.

We now understand the problem and have a solution. Let’s build that.

Peltarion with bubble

First, you will have to build an application specified in the use case. Still, wondering how your bubble app and peltarion experiment talk? It’s simple, we will use a plugin for their communication.
On your bubble app go to the plugins tab and install “peltarion”

We will install the first one. Now go to the workflow of your “add review button” and add an action “peltarion”.

You will find the required fields on your deployment tab in peltarion.

Put in your fields and in the “create a new review” add the key returned by peltarion for the sentiment.

Make sure to enable your deployment using the top right button “enable”.
Congratulations you have successfully integrated your pre-trained deep learning model in your application.

Working video

Link to app

From our blog

Latest Posts

Looking to build something?

Chat with our team to see what we can do