Over the past few months, AssemblyAI has been a proud sponsor at numerous hackathons around the world, including HawkHacks 2022 on May 21-22, 2022.
As a sponsor, we offered our Speech-to-Text API and Audio Intelligence APIs as resources for attendees to hack with during the event. We also provided real-time support for attendees, answering any questions they had about how to use the APIs or speech transcription in general.
Here’s a highlight of the top three projects built using AssemblyAI Speech-to-Text APIs during HawkHack 2022:
1. Speech2Code
Speech2Code is a web application that helps programmers write code easier and faster. Speech2Code lets users speak commands that they want to see implemented. Then, the web application translates these spoken commands into one of three written coding languages: Python, JavaScript, and Java.
The frontend of the application was built using React and the backend was built using a Python flask server for the AssemblyAI API integration, with Express.js for the code compilation.
In the future, the Speech2Code developers hope to expand the application to support additional languages and frameworks.
2. Vid2Text
Vid2Text is a web application that generates modified and customized audio and video transcriptions for uploaded audio and video files. With the application, users can automatically transcribe the files at high accuracy, modify the transcriptions as needed, and search and highlight keywords throughout the transcription text.
The application was built using Django, a Python web framework, and AssemblyAI Speech-to-Text and Audio Intelligence APIs for the transcription and keyword functionality.
In the future, the creators of Vid2Text would like to add multiple language transcription to the application and improve the overall user experience.
3. Vivy.ai
Finally Vivy.ai is an AI for Discord. With Vivy.ai, users can join a voice channel with the bot which will then verbally respond to anything the user says. The conversation is also transcribed in a text channel for users.
The creators of Vivy.ai used a basic discord bot to facilitate the AI’s development, as well as AssemblyAI for the Speech-to-Text transcription and GPT-3 to generate a contextual response.
In the future, the creators plan to create a 3D model of Vivy and have her “speak” and maybe even train her to sing.
