This demo uses WebRTC getUserMedia API for live microphone input. For the WAMI (Flash plugin required) version check the demo here (works in all browsers)
Configuration
Language and Acoustic model parameters:
N-best hypotheses:
Testing
Test recorded RAW audio:
Upload your own RAW/WAV (16bit PCM, little-endian), select corresponding sample rate (only 8kHz and 16kHz recordings supported due to the restrictions of the acoustic model) from Configuration and click Decode Audio:
Live microphone input (using WebRTC) -- enable Web Audio Input in chrome://flags
Select model from Configuration (sample rate is set by AudioContext object), click Record Audio, speak, and Stop | Decode