Rajasthani Multi-Dialect S2ST Studio

State-of-the-Art Universal Speech Translation & Zero-Shot Synthesis Engine

Whisper & Polyglot Online

Speech & Text Input

Text Mode Active

Recording length is unrestricted for this showcase. Long audio takes proportionally longer on the CPU server.

Voices run locally on our server. Expressive preview takes longer; regional pronunciation is still being evaluated. Listen to voice samples

Pipeline Results & Synthesized Speech

1
ASR Model
Waiting
2
NMT Translation
Waiting
3
Zero-Shot TTS
Waiting
Source (Marwari) ASR Transcription
Click "Run End-to-End Speech Translation" to begin...
Target (Mewari)
NMT Translation
Translated speech text will appear here...
Re-translate to:
Synthesized Audio Waveform Output 16kHz Mono WAV PCM
0:00 / 0:00
Audio Prep 0.00s
ASR Latency 0.00s
NMT Latency 0.00s
TTS Latency 0.00s
Total Latency 0.00s

Recent Translation History

No recent translations yet. Run S2ST pipeline to store session history.