Inspiration
Even though we are blessed to not have any mobility disabilities we tried to envision how we could interact with our smartphones by using only our sight and voice.
What it does
By starting from this premise we decided to develop a virtual assistant that will power the user interface. Some of the key points that we wanted from it would be to:
- need no touch events
- no difference in design to current implementation, (for example, the apple assistant adds small icons with numbers).
How we built it
-Behind the scene there is an API hosted in cloud that is called by the functions implemented in the code by developers, and this API process the data locally and with the help of other API: OpenAI's API.
Challenges we ran into
-In the beginning it is difficult to develop an intelligent assistant without other data in the database to have references and to compare the user's input.
Accomplishments that we're proud of
- Consistency across platforms: ios and android
- Easy to implement with as less as a line of text
- Be as familiar to the user as an analytics framework
- Be used a larger range of users, like commuting drivers.
What we learned
- We focused on cloud solutions to host our API that process the calls from the framework and get familiar with machine learning APIs that process data.
What's next for Virtual Assistant Framework
It can be trained without needing call external APIs besides our own.
Log in or sign up for Devpost to join the conversation.