What was made:
Vox Calculator — a voice-controlled calculator web app under Mash-Labs 01. It has a dark-themed UI, a mic button for voice input, a full manual keypad as backup, and it parses spoken phrases like “twelve plus eight” or “fifty divided by five” into arithmetic operations, plus voice commands for “clear” and “equals.”
What was likely challenging:
For a project like this, the hard parts are usually:
Parsing spoken number words (“twelve,” “fifty”) and operator words (“plus,” “divided by”) into a form a calculator engine can actually compute — this needs a decent grammar/parser, not just keyword matching.
Handling ambiguous or garbled speech recognition output (e.g. “for” vs “four,” “won” vs “one”).
Making voice and keypad state sync cleanly, so a spoken input and a tapped input update the same display without conflicts.
Browser speech recognition API quirks (permissions, mic access, inconsistent support across browsers).
“Prod” — I’m not sure what you mean here. Do you mean:
The production deployment (it’s live on GitHub Pages), or
The product (i.e., what it delivers to users), or
Something else specific to your build?
What people should know to test it:
Tap the mic button and grant microphone permission.
Say a full expression like “twelve plus eight” and it should compute automatically.
Say “clear” to reset, or “equals” to force evaluation.
Manual keypad works too if voice fails or isn’t available on the device/browser.
