This is a submission for the AssemblyAI Challenge : Sophisticated Speech-to-Text.
What I Built
Spext is a web app that helps international students study better when they are struggling to keep up with lectures because of the language barrier. Spext can
- Convert audio to text
- Translate audio to your language
- Convert and translate stream audio to text (Coming soon)
- Sumarize the audio (Coming soon)
- Ask questions about your audio data (Coming soon)
Demo
https://spext-ruby.vercel.app/
Journey
My technical stack is:
- NextJS for UI
- Firebase for authentication
- AssemblyAI’s Speech-to-Text Model APIs to convert audio to text
- An API to translate the language
- Vercel for hosting
I use Universal-2, AssemblyAI’s Speech-to-Text Model APIs to convert an audio to text
Then I create a feature which translates the converted text into the language that the user is fluent. Currently, I support 26 languages
In the future:
- I plan to develop a feature for real-time translation, allowing students to understand lectures as the instructor speaks instead of having to rely on recordings.
- I will support translation into more languages and make it more natural and human-like.
- I’ll create features to summarize audio, ask questions about its content, and generate flashcards based on the audio. <!-- Tell us how Universal-2, AssemblyAI’s Speech-to-Text Model was incorporated in your application. -->
Top comments (0)