Emapthica

Empathica is a personalized AI Social Good SaaS meant to make AI feel more human and give people 24/7 access to a mental health AI that knows them personally and helps them stay mentally healthy.

Comment

Inspiration I got the inspiration to build Empathica after seeing recent news articles about AI chatbots and systems being used more frequently to help people with their mental health. This hackathon presented the perfect opportunity to turn my interest into reality, so I took on the challenge to build out this SaaS.

What It Does Empathica is a personalized AI Social Good SaaS designed to make AI feel more human, providing people with 24/7 access to a mental health AI that knows them personally and helps them stay mentally healthy. The AI learns from the data you input and supports you through stressful times, serving as an always-available tool.

In addition to the AI chat feature, Empathica includes several sub-systems:

Health Tracker: Users can set daily goals, and the system creates a plan for them, allowing them to check off completed tasks. Guided Meditations: Offers 30-second to 5-minute guided meditations that users can access anytime. Journaling System: Users can write down their thoughts, struggles, and everything they are going through. Overall, Empathica is designed as a social good SaaS, using AI to assist people whenever they need it.

How We Built It I started by using a Next.js project with TypeScript and Tailwind CSS. I then incorporated a Prisma schema with an Azure SQL database for data storage.

Once the foundational database was set up, I created a user table and implemented NextAuth credentials for sign-up and sign-in functionality. Afterward, I focused on developing the core AI chatbot on the home page, allowing messaging and saving user threads.

Next, I built out the other sub-systems, utilizing APIs to handle backend communication, UploadThing for image storage, and other React-based tools for efficient development. Finally, I incorporated speech-to-speech interaction with the AI using OpenAI and Eleven Labs.

Challenges We Ran Into We faced several challenges:

Integrating OpenAI Whisper and Eleven Labs to manage audio files and responses promptly. Handling SQL table adjustments, as any small change required continuous edits. Ensuring that the website fit within different screen widths and heights. One of the most difficult parts was figuring out how to build something substantial in 20 hours while ensuring it looked polished.

Accomplishments That We're Proud Of I'm incredibly proud of how much I accomplished in such a short time. Many people consider me a fast developer, but I truly pushed my limits today in terms of speed and functionality. Successfully implementing the speech-to-speech feature was a major accomplishment, and I’m thrilled with the outcome.

What We Learned I learned that I’m capable of delivering a high-quality project in just 20 hours if I dedicate myself entirely. I picked up many new full-stack techniques and gained valuable experience with the Eleven Labs API, which I plan to use for future projects.

What's Next for Empathica I’m so enthusiastic about this project that I plan to pitch it at Red Bull Basement and see how it performs. I aim to further integrate the speech-to-speech model into the platform and develop a more dynamic UI for other services. My ultimate goal is to expand the AI's capabilities to make it even more human-like.

Built With

Share this project:

Updates