Inspiration:
We wanted to create a tool that makes coding more accessible and inclusive, especially for beginners and people who have difficulty typing. Voice2Code was born from this idea.
What it does:
Voice2Code converts natural voice commands into executable Python code using Whisper. Users can speak, view, run, and download the generated code within a friendly Streamlit app.
How we built it:
We used OpenAI's Whisper for speech recognition and Streamlit to build the interactive web interface. Commands are parsed using regex and mapped to Python code snippets.
Challenges we ran into:
Whisper’s output needed clean parsing for accurate command recognition. Handling ambiguous or unexpected speech inputs. Setting up real-time recording and processing on the frontend.
Accomplishments that we're proud of
End-to-end voice-to-code execution with a clean interface. Supporting multiple Python structures via voice. Making coding more accessible in just a few hours.
What we learned
How to integrate Whisper into real-time apps. Parsing natural language into code using regex.
Designing accessible, user-friendly interfaces with Streamlit.
What's next for Voice2code
Add support for more languages and frameworks. Improve command flexibility using NLP. Deploy it online and integrate with GitHub or VS Code.
Log in or sign up for Devpost to join the conversation.