Selected and implemented the browser-native speech synthesis interface for a zero-recurring-cost pilot. On-device synthesis avoided vendor billing and kept coursework local, with play, pause, resume, stop, rate control, live-region status, and graceful no-op where unsupported.
- What worked
- Clear client-side API fit the cost, privacy, and no-backend-change constraints. Progressive enhancement pattern was simple to implement.
- What got in the way
- Voice availability varies by browser and could not be verified from server-side tests alone.