HelpingHand

Inspiration

Helping Hand was inspired by our collective experiences with mental health challenges and the realization of the profound impact that supportive communities and resources can have on one's well-being.

We recognized the pervasive stigma surrounding mental health issues and the barriers individuals face in seeking help. Witnessing the struggles of loved ones and acquaintances navigating mental health challenges reinforced our belief in the importance of de-stigmatizing mental health conversations and providing accessible support systems. Helping Hand was born out of a shared vision to create a safe, inclusive online space where individuals could find resources and understanding.

What it does

1. Generative AI for Mental Health Support: HelpingHand utilizes AI to give immediate, confidential support to those struggling with mental health issues. Helping Hand offers an AI chat feature, designed to give users a place to voice their thoughts without having to speak to an actual person. This can be especially helpful for those with social anxiety who are nervous about contacting a hotline or other professionals but still would like a place to think through and vocalize their problems.

2. Affirmations Board: Users can write down affirmations on this highly customizable page which they can look back on when they need a helpful reminder to stay positive.

3. Gratitude Journaling: Gratitude journaling has been found to boost mood and positive thinking. Helping Hand uses AI to generate different writing prompts to guide users through this helpful process.

4. Resources Page: HelpingHand offers a resources page for those who feel ready to seek professional care, or learn more about mental health.

5. Donations Page: Users who would like to give back to mental health related organizations have an area where they can seek out the organizations and contribute.

The ai chat, affirmation board and gratitude journaling pages save their data locally and are not stored in any online database, giving users peace of mind while still providing the ability to save data across sessions.

How we built it

1. Frontend Development With React: We utilized React.js to build the highly interactive and visually pleasing frontend. React was crucial, as it allowed for easy connection to backend servers, and excellent ability to create a seamless user interface.

2. Backend With Express: Express.js was our choice of backend technology, since we could easily create a connection to the React frontend. Express was where we handled all the communication with AI services.

3. Node for Package Management: We used Node.js for package management due to a variety of open source projects which we could add to our project to increase the functionality and user experience.

4. Generative AI With OpenAI: We used OpenAI api to handle all of our generative ai tasks, from automating the creation of unique gratitude questions, to the user chat functionality.

Challenges we ran into

While developing Helping Hand, we ran into a few challanges that really tested our problem solving skills:

1. Linking front end and back end: Initially we used fastapi with uvicorn to create a local server that would interface with the openai api. Unfortunately the server would only work correctly on one of our computers, requiring us to quickly pivot to a different system. We decided to use Express.js to create our backend server which, after a bit of work, ended up functioning properly allowing us to stay on schedule.

2. Audio processing: We wanted the ai chat to be as conversational as possible, so we planned to implement speech-to-text and text-to-speech functionality to more accurately simulate talking to someone. This ended up being much more complex than we anticipated and led to lots of setbacks.

3. Use of unfamiliar technology: Most of our team are very new to technologies such as React, Express, and OpenAI, so learning these new technologies quickly and creating a functional app was a very challenging process.

Accomplishments that we're proud of

1. Successful Usage of OpenAI API: The integration of generative AI into HelpingHand was an amazing accomplishment, since every user can get an interactive, unique experience which is available to everyone at all times.

2. Saving User State Confidentially: The ability for our program to save the program state for each user without needing a database is something we are very proud of, since confidentiality for mental health topics is very important. This feature also allows for a much more interactive experience, allowing the user to revisit pages such as the affirmation board when they need a mood boost.

3. Creation of Interactive User Pages: Interactive, useful, and intuitive user interaction is something we did very well in HelpingHand, since users can create new elements, interact with them, and achieve satisfying results.

What we learned

1. AI Technology Support: Integrating AI into our application let us gain lots of insights into how AI models work, and how to optimize models based off of prompts to produce favorable results.

2. Coding Best Practices: HelpingHand was a very large project, containing about 16 files which us developers worked on. In order to maintain a clean, working codebase which is readable and functional, we had to make sure to follow proper coding style, and software design principles.

3. Adaptability: During this project we faced lots of setbacks, and had to pivot to using new technologies, which helped us think on our feet, and generate new code quickly, and accurately.

What's next for Helping Hand

1. Enhanced AI Model: While Helping Hand currently uses a standard general purpose ai model, we would like to eventually develop a more specialized version that is even better suited for conversation and emotional support.

2. Interactive De-stressing Page: When the user is stressed out, we would like to add a page which could help them remain calm, and improve their mood.

3. Voice User Interaction: We would like for user audio processing speech-to-text, where the user can talk into their mic and receive a response.