Voice Transcription

Control voice input across all Elikias platform interfaces

Global Voice Input

Master switch for all voice input across the platform.

Voice input is ENABLED

Affects all scopes below. Disabling this removes all mic buttons platform-wide.

Transcription Mode

How voice input is transcribed. Web Speech is free and default.

Web Speech Only

Default

Uses browser-native speech recognition. Free, zero-latency, no server calls. Limited to ~50 languages.

Hybrid Mode

Recommended

Uses API credits

Web Speech API primary. Falls back to OpenAI Whisper for unsupported languages (Lingala, Kikongo, Yoruba, etc.).

Whisper Only

Premium

Uses API credits

All transcription routed through OpenAI Whisper backend. Highest quality and language coverage. Uses API credits.

Backend transcription is OFF. Only browser-native Web Speech API is used.

Recording Settings

Max recording duration: 60s

15s5 min

Usage logging

Log voice transcription events for review

Scope Controls

Enable voice input per interface area. Disabled by default (except Assistant).

Elikias Assistant

AI health chatbot voice input

Community Chats

Community posts, comments, messaging

Lisanga

Meeting chat composer

Role Messaging

Patient/pharmacy/physician/pharmacist DMs

Admin Tools

Admin chat and support tools

Provider Information

Web Speech Provider

Browser Native (free)

~50 languages, varies by browser

Backend Provider

OpenAI Whisper

Model: whisper-1 · 99+ languages

Backend transcription (Whisper) uses your OpenAI API key and incurs costs (~$0.006/min). Enable only when needed. The OPENAI_API_KEY is securely stored as a backend secret and never exposed to the browser.

Admin Notes

Elikianet

Voice Transcription