people-and-co

TherapyAI: Your 24/7 Pocket Therapist—No Couch, No Judgement, Just Support.

Comment

Inspiration

One of our team mates was finding difficulty in finding affordable and quick therapeutical care during a crisis, this gave us the idea on how we should build something that can be accessed on the spot and free of cost.

What it does

  • Conversational AI: Engage users in empathetic, supportive, and non-judgmental conversations.
  • Mental Health Support: Provide coping strategies, resources, and active listening.
  • Memory Retention: Retains context from previous conversations for coherent and context-aware responses.
  • User-Friendly Interface: Simple and accessible web-based interface.
  • Privacy-Focused: No data storage or tracking of sensitive user information.
  • Free to Use: Built entirely with free tools and APIs.

How we built it

Backend: Built using a server-side framework (Python) to handle logic and API (Gemini 1.5 Flash). Model was fine-tuned using a large dataset pulled from a verified source. Features included RESTful endpoints, authentication (JWT/OAuth), data validation, server deployment, and memory retention. Testing was done with tools like Postman/unit tests.

Frontend: Developed with a client-side framework (React) for UI/UX. Included responsive design (HTML/CSS), state management (Redux/Context API), and API integration (Axios/Fetch). Deployed via static hosting (Vercel) and tested.

Integration: APIs connected frontend/backend, ensuring secure communication (CORS, HTTPS). Git/GitHub managed version control, with CI/CD pipelines automating testing/deployment.

Challenges we ran into

Major challenge we ran into was successful integration of frontend and backend parts of our code.

Accomplishments that we're proud of

We were able to make our idea into something that has great potential, and can be scaled in the future with proper funding.

What we learned

Apart from the coding, we learned a lot ranging from version control, system design, databases, testing, fine-tuning, API collection, model refinement, cloud technology etc.

What's next for people-and-co

We plan to scale this operation in such a way that will allow users to store their private data into a profile that they can visit anytime late. We plan to add features such as group therapy sessions in which many users that are in the same boat can virtually connect and have a virtual therapist AI bot preside over.

Built With

Share this project:

Updates