LLM Voice Interaction

Configure, interact, and review your LLM conversations.

Setup

Recognition Language (STT)Affects STT recognition and default TTS mapping.

Advanced: Manually choose TTS voice

Enable voice override

When disabled, the app auto-selects the best voice for the chosen STT locale.

Current STT Locale: English (United States)

Current TTS Voice: Auto (pending)

TTS voices are not available in this browser. The app will operate in text-only mode.

Recognition language: English (United States)

This browser does not support Speech Recognition. Use the text fallback below.

Interim transcriptText fallback / Final prompt

Model: gemma-3-12b-it

Prompt

This browser does not support speech synthesis. You can still use text responses.

Recognition language (STT): English (United States)

Voice: Auto (best match)

Rate (1.00)Keep within 0.7–1.6 for intelligibility.Pitch (1.00)Safe range 0.6–1.4 to avoid artifacts.WPM (150)Affects fallback time-slice pacing.

When voice language does not match the STT localeTranslation uses the existing proxy and counts toward your quotas.

Text to speak

Model response will appear here.

Status: unknown

Use this until STT/TTS are wired. It exercises Dexie + LRU policy.

PromptResponse

Locale: en-US