Nostalight

Inspiration

Everyone has to face the time when they lose someone who is dear to them. It is not easy to cope with the pain of losing loved one and people need time to cope with it. Thus, we would like to build an app that can help people find peace and relief by letting chatbot pose as people's loved ones and console people to overcome the losing pains.

What it does

Nostalight gathers information about name, hobbies, characters of the lost loved one of the user and creates a chatbot that represents the lost loved one. The chatbot will pose as the pass away loved ones, talk to the users, and help them overcome the pain of losing their loved ones.

How we built it

AWS PartyRock is used as the platform to host our front end and back end app. Model Claude (LLM) was used for text analysis and generate context as well as advice, dialogues for the chat bot to talk to the users. AWS PartyRock allowed us to create user input variables that intakes traits of the lost ones. All data are fed into the system to create the chatbot that mimic that lost loved ones and allows the chatbot to pose as the lost loved ones and talk with the users.

Challenges we ran into

We had challenges to learn and come up with idea that is useful for humans' lives. We had challenges building the dialogues for the chatbot to response to the users who lost their loved one. We try to make our chatbot mimic the characters and personalities of the lost people as closely as possible.

We had challenges of trying to put ourselves in the shoes of people who lose someone that are dear to them and try to make the conversation as meaningful and full of love as possible.

Accomplishments that we're proud of

We were able to build an app that helps people who lose loved one to feel better through talking with their loved ones and get consolations. The app showcases the power of generative models and natural language processing which are used to build the app. The app is complicated and was built based on long thought process and UI/UX consideration to intake data and create a chatbot that can mimics that lost loved one as closely as possible.

What we learned

• We learnt to build the app using PartyRock AWS, LLM, and Claude. We read documents and learned a lot about generative AI as well as LLM.
  
• We learned about the UI/Ux process of how to put yourself in the thought process of the users and figure out which functions or input are needed for creating the chatbot mimic the lost loved ones.
• We also learned to manage time to work and deliver the final project on time.

What's next for Nostalight

We want to add more advanced functions to the app such as:

• Allow users to upload photos of the lost loved one
• Users can upload voices of the lost loved one for the chat to mimic
• Create live image of the lost loved ones that can talk with the users directly instead of typing
• Having live voice input and output for the chat that makes users feel the conversation is real and feel immersed in the conversation
• Allow the ability to create multiple chatbots pose as multiple lost loved ones to talk to the users

Built With

• amazon
• amazon-web-services
• claude
• llm
• partyrock