Start talking
Billable operation: Deepgram realtime STT begins only when you press Start Live Mic.
Start Live MicCtrl Enter StopShift Esc ClearR
Choose a microphone, then start a live token-based Deepgram session.
Live transcript
Your transcript will appear here. Last successful state
None yet
Microphone Input Default microphone Choose your condenser mic or audio interface if it appears here.
Selected: Microphone 1
Refresh Devices Realtime language configuration
This setting tells Deepgram which language you will speak. It does not translate speech into another language.
Recognition mode Known spoken language Use one verified Nova-3 streaming language.Nova-3 multilingual / code-switching English, Spanish, French, German, Hindi, Russian, Portuguese, Japanese, Italian, Dutch.Spoken language English (en) Italian (it) Spanish (es) French (fr) German (de) Portuguese (pt) Japanese (ja) Dutch (nl) Arabic (ar) Arabic (United Arab Emirates) (ar-AE) Arabic (Saudi Arabia) (ar-SA) Arabic (Qatar) (ar-QA) Arabic (Kuwait) (ar-KW) Arabic (Syria) (ar-SY) Arabic (Lebanon) (ar-LB) Arabic (Palestinian Territories) (ar-PS) Arabic (Jordan) (ar-JO) Arabic (Egypt) (ar-EG) Arabic (Sudan) (ar-SD) Arabic (Chad) (ar-TD) Arabic (Morocco) (ar-MA) Arabic (Algeria) (ar-DZ) Arabic (Tunisia) (ar-TN) Arabic (Iraq) (ar-IQ) Arabic (Iran) (ar-IR) Belarusian (be) Bengali (bn) Bosnian (bs) Bulgarian (bg) Catalan (ca) Chinese (Cantonese, Traditional) (zh-HK) Chinese (Mandarin, Simplified) (zh) Chinese (Mandarin, Simplified - China) (zh-CN) Chinese (Mandarin, Simplified Han) (zh-Hans) Chinese (Mandarin, Traditional - Taiwan) (zh-TW) Chinese (Mandarin, Traditional Han) (zh-Hant) Croatian (hr) Czech (cs) Danish (da) Danish (Denmark) (da-DK) English (United States) (en-US) English (Australia) (en-AU) English (United Kingdom) (en-GB) English (India) (en-IN) English (New Zealand) (en-NZ) Estonian (et) Finnish (fi) Flemish (nl-BE) French (Canada) (fr-CA) German (Switzerland) (de-CH) Greek (el) Gujarati (gu) Gujarati (India) (gu-IN) Hebrew (he) Hindi (hi) Hungarian (hu) Indonesian (id) Kannada (kn) Korean (ko) Korean (South Korea) (ko-KR) Latvian (lv) Lithuanian (lt) Macedonian (mk) Malay (ms) Marathi (mr) Norwegian (no) Persian (fa) Polish (pl) Portuguese (Brazil) (pt-BR) Portuguese (Portugal) (pt-PT) Romanian (ro) Russian (ru) Serbian (sr) Slovak (sk) Slovenian (sl) Spanish (Latin America) (es-419) Swedish (sv) Swedish (Sweden) (sv-SE) Tagalog (tl) Tamil (ta) Telugu (te) Thai (th) Thai (Thailand) (th-TH) Turkish (tr) Ukrainian (uk) Urdu (ur) Vietnamese (vi) All languages supported by Nova-3 and the realtime /v1/listen transport. Language support varies by model, endpoint, and transport.
Try your native language
English Italian Spanish French German Portuguese Japanese Dutch Hindi Thai
Recognition mode Known spoken language Model nova-3 Spoken language English Language code en Endpoint WSS /v1/listen Audio MIME detecting... Interim / endpointing / VAD on / 300 ms / on Auth Temporary token
Billable operation: Deepgram realtime STT begins only when you press Start Live Mic.
Optional native-language presenter prompt Hello, I am testing realtime transcription. I will pause, then correct one detail.
Flux multilingual: Available in the Applied Voice Systems turn-taking lab. The current Live Mic keeps its stable Nova-3 v1 lifecycle; Flux v2 requires different turn events and capture timing.Open Flux Lab
Official model and language reference Transcript redaction
Changes transcript output only. The original audio remains unchanged.
Open in Redaction Lab Redaction policy Off General PII Payment Data Health Information Numeric Identifiers Aggressive Numeric Masking Financial Contact Center Healthcare Contact Center Off — no redact query values will be sent.
Local mic input level 0%
The meter only confirms local microphone input. Deepgram transcription starts after the WebSocket opens.
Selected browser MIME type detecting...
Deepgram connection
Token
WebSocket
Recorder
Transcript events
Refresh Devices Start Live MicCtrl Enter StopShift Esc Clear TranscriptR Run 5-sec Mic Test
Run 5-sec Mic Test
Choose a microphone, then start a live token-based Deepgram session.
Live audio uses two bounded connection attempts. Containerized browser audio is sent without raw PCM parameters.
Live mic diagnostic checklist
API key detected pending
Temporary token route works pending
Microphone permission granted pending
Input device selected pending
Browser MIME type supported pending
WebSocket opened pending
Audio chunks sent pending
Deepgram events received pending
Final transcript received pending
Interim Transcript
No interim transcript yet. Final Transcript Log
No final transcript yet. Configured spoken language: English (en). This is configured, not detected.
Raw live event JSON [] Demo script to read aloud
SaaS webhook issue
The webhook payload for customer ID CUS-1842 failed twice, then the retry policy sent a duplicate event to the dashboard.
Italian account support
Buongiorno, non riesco ad accedere al mio account e vorrei controllare la fattura numero IT-4821.
Interview intro
I built this Deepgram learning lab to test live microphone transcription, sample audio, and text to speech from one local dashboard.
