January 5, 2017

Clarifai Featured Hack: Improve Your Emotional Health With Feelybot

Table of Contents:

It’s a new year, which means everyone is refocusing on what’s important to them, whether it’s getting back in shape or taking a year off to travel. One thing we believe to be super important is emotional health – that’s why we love FeelyBot, an app that senses your mood and improves it through positive interactions.

FeelyBot is an app that aims to be your closest and most loyal friend; he will always be there for you through the good and the bad. His sole purpose is to improve your day through positive interactions via the “super cute” Jibo platform and, in the long term, help treat behavioral and emotional disorders like depression and social anxiety.

Through a user’s social media accounts, FeelyBot can evaluate the strength of the user’s five basic emotions: happiness, sadness, anger, disgust, and fear. It then uses that information along with a contextual understanding of text and images to interact with the user and improve their mood!





When you’re feeling down, getting affirmations from FeelyBot is a lot more effective and healthy than eating your feelings. Just saying. Read more about FeelyBot on DevPost or check out the GitHub repo!


We caught up with Minh Hoang, one of the creators of FeelyBot, to talk about how his team built the app that gives us all the feels.

Clarifai: What inspired the idea for FeelyBot?

Minh: I like to take pictures and I figured that it would be really cool to have a timeline of all my photographs based on the feelings that they captured. Then, we figured that we could use that information to help people with emotional disorders. There are millions of people in the world that suffer from depression, social anxiety, or other forms of disorders. We asked ourselves how we could use the Jibo platform and it’s super cute interface to cheer up or improve a user’s mood.

How did you build the app?

We used JavaScript pretty much for everything. The server will track if there is a new Facebook post and run a sentimental analysis on the post. There are two types of post that are going go to be analyzed: text posts and image posts. With text post, FeelyBot will call an AlchemyAPI, which scores the different emotions of the post. FeelyBot then interprets the result and responds to the user in a positive way. With image post, FeelyBot will collect the image, then call Microsoft Emotion API to analyze the sentimental elements and Clarifai to get an overall description of the image. FeelyBot combines the results from both APIs and uses it to determine the user’s emotional state. All API is being called via Rapid API or native API.

The Jibo platform POSTs information to the Node server in order to get the processed information and interact with the user through text-to-speech. The user’s input is then sent back to the Node server, where it is being constantly analyzed and used to improve the backend’s AI to make it more personalized.

What was the best part about working with the Clarifai API?

Clarifai API is very friendly to use. Thank to the Clarifai team, all documents for the API are straightforward and easy to read.

Thanks for sharing, Minh!

To learn more, check out our documentation and sign-up for a free Clarifai account to start using our API – all it takes is three lines of code to get up and running! We’re super excited to share all the cool things built by our developer community, so don’t forget to tweet @Clarifai to show us your apps.

And give Minh some props in the comments below. Until next time!