Inspiration
We wanted a way to help others get more comfortable with public speaking and develop "soft" interview situations
What it does
Analyzes speech metrics and transcript, evaluates the answer and provides descriptive feedback
How we built it
OpenAI API to get sample questions and feedback to users HTML, CSS, JavaScript frontend C# backend FFmpeg to extract audio and whisper-1 for speech-to-text
Challenges we ran into
Getting AI to return a strict JSON AND descriptive feedback to was pertinent to each individual situation Effectively evaluating each performance Learning new languages
Accomplishments that we're proud of
Learning new languages Built a functional project in 24h
What we learned
Utilizing APIs Prompt engineering Keeping a clean architecture in the backend
What's next for SpeakEasy
Attention recognition (camera) Code evaluation More metrics More real-world applications
Built With
- ai
- api
- c#
- css
- html
- javascript
- openai
- typescript
Log in or sign up for Devpost to join the conversation.