v0.9 — Voice in 27 languages
Voice interface for the AI chat widget, available in 27 languages across all plans.
Release date: 1 April 2026
What's new
Voice interface
Visitors can now speak to the elNudge AI agent by tapping the microphone button in the chat widget. The agent listens, processes speech to text, generates a response, and speaks the response back to the visitor — all within the chat widget. No additional setup is required on the visitor's side.
Voice is opt-in per site. To enable it, go to Dashboard → Settings → Voice and toggle voice on for the relevant site.
27 supported languages
Voice input and output are supported in all 27 languages that elNudge supports, including:
English, Hindi, Tamil, Telugu, Kannada, Malayalam, Bengali, Marathi, Gujarati, Punjabi, Urdu, Odia, Assamese, Sanskrit, French, German, Spanish, Portuguese, Arabic, Indonesian, Japanese, Korean, Mandarin Chinese, Thai, Vietnamese, Turkish, Swahili.
Language detection is automatic — the AI responds in whichever language the visitor speaks.
Voice quality scales by plan
| Plan | Voice quality |
|---|---|
| Starter | Browser-native text-to-speech (quality varies by OS and browser) |
| Growth | Higher-quality AI-powered neural TTS |
| Scale | Premium neural voices with lower latency |
| Enterprise | Custom voice configuration available |
All plans use the same Deepgram speech-to-text engine for voice input.
Firefox fallback
Firefox does not support the browser microphone API required for voice input. On Firefox, the microphone button is automatically hidden and visitors interact via text chat. No configuration is needed — the fallback is handled by the SDK.
Breaking changes
None. This release is fully backward-compatible.
Migration
Voice is opt-in. Existing sites are not affected until you enable it in the dashboard. To get started:
- Go to Dashboard → Settings → Voice.
- Toggle Enable voice on.
- Optionally, set a monthly voice cap to control overage.
See Voice Minutes & Overage for details on included minutes and overage rates.