Lovely

Dashboard
Automated messaging

Inspiration

Social battery? Drained. Relationships? Nonexistent. Time to commit to a talking stage that will likely get to nowhere? None.

This is what a lot of people are experiencing, and it is quite suboptimal.

Finding love should be easy, not stressful and near-impossible, and there is nothing better to ensure a comfortable dating experience than the very tool we put all our trust into: AI.

Using AI, we wanted to revolutionize the dating scene and increase peoples' chances of finding a partner (and efficiently).

What it does

Lovely facilitates the dating scene for busy individuals by automating the texting stage, video calls, and even real-life, face-to-face interactions. It engages in back-and-forth messages on the user's behalf through WhatsApp, and uses a facial scan of the user's face along with a sample of their voice to mimic an authentic video call with the person of interest. Lovely ensures a strategically romantic tone while conversing with this person and can set up dates, then notifying the user of how and where to attend the date. During the date, the user can use Ray-Ban Meta Glasses to further benefit from Lovely, which gives real-time feedback to the user through the glasses' speakers.

How we built it

Our backend combined Python and TypeScript, with Supabase handling database and login. We used the WhatsApp Business API to send messages with a conversational tone using a custom logic layer that tracks chat history. For video calls, we generated AI avatars with Heygen and used Twelvelabs to understand when to escalate. The Meta Smart Glasses gave live feedback during real dates, while the frontend let users log in, view contacts, and see date insights using Figma.

Challenges we ran into

We initially planned on using Instagram Direct Messaging for the text conversations; however, we switched to WhatsApp because it provided a more seamless experience with the API.

Accomplishments that we're proud of

Combining so many popular aspects of AI to create a generational idea
Finally implementing these Meta Glasses into a coding project of our own
For 1/2 of our team, starting off our hackathon careers with a blast!

What we learned

We learned how to work with the Meta Smart Glasses: setting up Bluetooth, sending voice prompts, and syncing feedback in real time. We also figured out how to coordinate several APIs and tools to work together across different stages of interaction, which helped us better understand multi-platform app architecture.

What's next for Lovely

A potential future implementation of Lovely would be to incorporate a visual coaching aspect into real-life dating interactions, ideally through the lens of the Meta Glasses.