Bee Voice-Dev Assistant

🐝 Inspiration

As open-source developers and students, we frequently encounter bugs, visual glitches, and runtime errors while exploring repositories. Reporting them manually—stopping development, opening GitHub, writing structured issues, and collecting logs—creates friction that often leads to unreported bugs. We wanted to build a seamless solution using voice-first AI and modern developer tools.

🚀 What it does

Bee Voice-Dev Assistant allows developers to report bugs and terminal errors using voice input. The application transcribes speech, processes it via AWS Bedrock (Claude model) to generate structured issue drafts, and publishes them directly to GitHub via API, complete with a Socratic AI mentor for debugging guidance.

🛠️ How we built it

• Frontend: React 18, Vite, Tailwind CSS, and browser Web Speech API for real-time voice recognition.
• Backend: Node.js, Express.js, TypeScript, and Socket.IO for real-time event communication.
• AI & Integration: AWS Bedrock Runtime SDK (@aws-sdk/client-bedrock-runtime) and GitHub REST API.

🚧 Challenges we ran into

• Handling browser microphone permissions and cross-browser Web Speech API compatibility.
• Configuring secure IAM roles and permissions for AWS Bedrock model invocation on the backend.
• Structuring real-time socket events between the React client and Express server without dropping state.

🏆 What we learned

We gained deep experience integrating multi-modal voice inputs with enterprise cloud AI (AWS Bedrock) and automating developer workflows securely.

Built With

• ai-apis
• amazon-web-services
• aws-bedrock
• claude
• cloud
• css3
• express.js
• full-stack
• git
• github
• github-api
• html5
• javascript
• mongodb
• mongoose
• node.js
• react
• rest-api
• socket.io
• tailwind-css
• typescript
• visual-studio-code
• vite
• web-speech-api
• websockets