Voice-Controlled Crossing Alert Bot

I want a lightweight web application that listens to a user’s voice command—think “Is it safe to cross?”—analyzes live traffic data, and immediately alerts when it’s the right moment to step off the curb. The only must-have feature for this first release is the real-time “alert when to cross,” but the architecture should stay clean enough that I can layer in richer predictions later. Here’s how I picture the build: • Voice interface. A browser-based speech-to-text component (Web Speech API, Vosk, or any comparable library) that wakes on a hotword or button press, captures the question, and triggers the analysis. • Traffic data hook. Query an open or paid traffic flow API, or process video feed input if that’s simpler, then distill the feed into a go / wait decision. • Alert logic. When conditions are right, return an on-screen message and an audible cue; otherwise tell the user to hold. • Clean deployment. Hostable on a standard Linux server—Node.js, Flask, Django, or another widely supported stack is fine—bundled with a concise README and environment file so I can redeploy fast. Deliverables 1. Fully commented source code in a public or private repo. 2. Hosted demo URL for acceptance testing. 3. Step-by-step setup guide (install, environment variables, API keys, run). If you have prior projects involving voice UIs, real-time data streams, or pedestrian safety tools, feel free to share links; they’ll help me gauge fit.