ValiAC

• 

  10000+ image large language model

Inspiration

We wanted to create an app for the LGBTQ community because the lack of resources and education provided is a very prevalent problem in the world currently, with a lot of bigotry towards the community even being enforced by law in some countries. This is important to us personally as well, as many people we know, including Anton himself, are part of the community. We hoped to be able to fix a major problem for LGBTQ people and came up with the idea of ValiAC by combining our extensive knowledge of well-being and self-growth psychology with our desire to help others.

What it does

ValiAC’s main purpose is to make people feel comfortable in their environment and know that they are supported. By providing three different ways in which users can do so, we ensure that people who feel supported and comfortable in different ways are still able to use the app. The most innovative implementation of this idea comes from the self-reflection tool. In it, users can quickly add their current mood and look back on it whenever they want. B.J. Fogg, the author of Tiny Habits, refers to actions that are easy to do and provide great benefits as Golden Behaviors. After refining the Mood Board for quickness (on paper), Swastik also tried it out for a week and found that the combination of the speed at which he was able to track his mood and the benefit that he could figure out trends in his emotions was extremely useful in daily life–a Golden Behavior. Our personal chat feature is a generative AI model that is powered with sentiment analysis tracking and combines speech-to-text and text-to-speech to create a situation in which people can text or speak into the mic and the app will respond in a supportive and reassuring manner. Finally, our community chat feature is simply a place where people can speak about anything on their minds anonymously and engage in conversation with others.

How we built it

The app is comprised of a Flutter frontend and a Google Firebase backend. The backend stores login information in Cloud Authentication and uses Cloud Firestore to store data. We used Firebase as it is a fast, reliable, and cheap service that runs on a very generous freemium model. Next, we built the machine-learning model to detect emotions using Google's Teachable Machine and connected it to Cloud Machine Learning so that it could easily be deployed. we also used a variety of Flutter APIs and speech transcription libraries to build our personal conversation bot and community chat, including Chat GPTs generative AI, sentiment analysis builder, speech-to-text/text-to-speech, etc. We combined ChatGPT with sentiment analysis, speech-to-text, text-to-speech, and a few other libraries in order to create a bot that you can fully converse with through voice or chat and it will provide information about a variety of topics related to mental health or the LGBTQ community or support/reassure you.

Challenges we ran into

Coding this out was extremely challenging. We had to meticulously work through bugs and errors for hours and hours to complete this. We spent the entire time fully focused and it definitely paid off.

Accomplishments that we're proud of

We're really proud of how much we've done with this app. There is an unbelievable amount of work that went into this project and we're just really glad it all worked out.

What we learned

Having a purpose and constantly pursuing that purpose gets you very far very fast.

What's next for ValiAC

The community chat is one to which we plan on adding daily checks to prevent bad-mouthing, misinformation, and hate speech. There are definitely tons of improvements in the emotion detection feature and the responses for the chatbot, but we're still very glad we got this far, too.

Built With

• flutter
• gpt
• llms
• vision