Levity
Levity allows you to automate repetitive tasks through the power of machine learning. Without writing a single line of code, it’s possible to train, deploy, and integrate an AI model within your Bubble application.
If you’re working in Bubble, you’re more than likely familiar with the process of creating workflows. These rule-based actions allow you to visually streamline repetitive, static tasks within your app.
But how can you automate tasks that still require human input? For example, how can your Bubble app determine what an image of a cat looks like? Or how can it even analyse the sentiment of a user-generated comment
Levity gives you the power of cognitive workloads, without the complexity of any code. Their platform allows you to build powerful AI models that are capable of making human-like decisions. This can streamline tasks that require a detailed level of judgement.
What Can You Build With Levity
Without having any prior experience as a data scientist or engineer, Levity makes it possible to create a fully-fledged AI model. These models are capable of understanding the lexicons of text, or even the objects within an image.
Levity’s core strengths lie in building AI for documents, images, and text, that automate your daily repetitive tasks. Some key use cases can include;
- Creating a model to analyse the sentiment rating of text-based data like customer reviews or tweets.
- Building an image classification model that can easily determine what objects are within a picture.
- Classifying the contents of a PDF document.
Levity even provides a great list of use cases and pre-vetted datasets to help you get started.
Getting Started With Levity
Getting started is as easy as creating an account, then activating your 14-day free trial. Without any previous engineering experience, I was able to quickly train and deploy my own model in less than an hour.
Once you open an account, you can create your first AI block, then upload the data set you wish to train the model on.
From here, training your model is as simple as adding custom labels to each of your dataset entries.
In the example below, you’ll see I’m labeling each of my car images based on the type of damage it’s received.
Finally, after your model’s been trained, you can test a working version of it directly inside Levity, then connect it to your Bubble app through their plugin.
