Text-to-speech runs entirely in the frontend (rise-dc-app/src/shared/utils/textToSpeech.ts), pulling VITE_SPEECH_KEY and VITE_SPEECH_REGION into client code and exposing credentials.
Move speech synthesis calls to a backend route that calls the Microsoft Cognitive Services APIs; frontend should call this backend and receive audio blobs/URLs. Remove direct SDK usage and secrets from the Vite bundle and load keys/region on the server only.
Deliverables: No speech keys in client builds, speech still works via backend endpoint, updated README with required server-side env variables.
Text-to-speech runs entirely in the frontend (
rise-dc-app/src/shared/utils/textToSpeech.ts), pullingVITE_SPEECH_KEYandVITE_SPEECH_REGIONinto client code and exposing credentials.Move speech synthesis calls to a backend route that calls the Microsoft Cognitive Services APIs; frontend should call this backend and receive audio blobs/URLs. Remove direct SDK usage and secrets from the Vite bundle and load keys/region on the server only.
Deliverables: No speech keys in client builds, speech still works via backend endpoint, updated README with required server-side env variables.