DAME

DAME is a speech-to-speech assistant with the purpose of providing users a one-stop-shop to adventure.

Comment

image

Inspiration

The inspiration behind DAME was the frustration associated with planning a trip. We know that travel agents aren't accessible to everyone and that shopping for different elements of your trip, from flights to hotels, can be frustrating. We have all been there: twenty tabs open, looking for the best deal. But DAME is here to change that.

What it does

DAME is an intelligent speech-to-speech bot for planning your travel itinerary. DAME prompts for a few pieces of information about your trip and automatically generates an itinerary for your entire trip. DAME utilizes Microsoft Semantic Kernel with API integrations to provide a truly customized and easy travel planning process.

How we built it

We built DAME using a React Native front end with the Whisper API speech-to-text and ElevenLabs text-to-speech. Using Whisper and ElevenLabs together allows for a speech-to-speech interaction that feels very natural. In the backend, Microsoft's Semantic Kernel orchestrators process the incoming information using GPT-3.5-turbo and infer parameters to pass to our custom semantic plugins to pull information from Google Places API, Ticketmaster API, and Amadeus API. This information is further processed by the semantic kernel to form an itinerary for the user. Because of our API integrations, DAME can plan flights, hotels, national park visits, restaurant visits, concerts, food & wine festivals, and more!

Challenges we ran into

Learning and utilizing Microsoft Semantic Kernel was a BIG challenge for us. It is relatively new, which makes troubleshooting more difficult. Our team also had two first-time hackers who had to persevere through a steep learning curve of cutting-edge technologies and learned a ton.

Accomplishments that we're proud of

Completing the project on time is a significant accomplishment for our team. There were times when we had run into complicated bugs that made finishing in time feel unlikely. Luckily, through continued perseverance, team work, and a lot of caffeine, we were able to overcome all of our issues to present a working demo.

What we learned

Our team has a wide range of skill levels and everyone walked away learning something new. From learning API integrations, to the semantic kernel, there were new things for everyone on the team to tackle.

What's next for DAME

Right now, DAME can only make the plans for you and cannot actually book any part of your trip. In the future, we would like to allow DAME to book the entirety of your trip for you so you can eliminate all frustration from your trip shopping experience.

Built With

Share this project:

Updates