MeditationAI

Everything is our modern day world is so fast pace, but we don't take the time to stop and discover how we truly feel. The ancients have long discovered meditation, and now we are brining it back w AI

Comment

Inspiration

The inspiration for this was to create to help us better understand ourselves. It allows us to create a world in which we answer questions, along with receiving wisdom along the way, to find inner-truth.

What it does

This application builds a scene where the user answers a prompt, and then the application, using AI will generate a response to the question. The AI will determine the mood of the response, and by extension, the user, to create an image and a selection of music to create the ideal environment for growth.

How we built it

We build the UI using Godot. And then for the backend, we used Ollama for the main LLM duties and Huggingface for the Image creation.

Challenges we ran into

The biggest challange we ran into was the creation of the visualization. We wanted to use projectM to dynamically create visuals what went with the audio. However, this library proved to be very difficult to use as it lacked an easy to use API. We decided to instead generate an AI image, and in this way, it could be in theme with the answers that the user gives.

Accomplishments that we're proud of

The sync of having all these different AI pieces come together; the synergy of the music, image background generation, and the adaptive prompting to the user is really rewarding. Our LLM prompts to the user a meditative experience in which the user can select how they are feeling and choose to proceed there. Our LLM creates a description to help generate an image based on this experience which is displayed to the user and read allowed (in which we added reverberation to amplify the euphoric experience), as well as helping choose from a plethora of music based on its intensity, and then proceeds to go further into the journey in an almost "pick your own adventure" style. Each choice you make as a user impacts the next coming prompt, experience, background, and music.

What we learned

We learned the strengths and disadvantages of AI. While it can be very powerful, it requires careful tuning of it to ensure that the output is correct while staying healthy and non-offensive to the users. Also, we learned of how to make all these different AI models come together to create a cohesive experience. However, more importantly, we learned more about mental health and inner-peace.

What's next for MeditationAI

What's next? It's just the beginning! In inspiration of the short-term videos, we would like to have text appear on screen instead of just being read aloud, as studying show that it increases attention and memory. Adaptive poses of the figure on screen, this would help us be more interactive with the user and create a sense of trust. Lastly, having different speakers would be a big help to help the user pick a narrator that the identify and connect with.

Built With

Share this project:

Updates