Baidu Speech Recognition

Baidu Speech Recognition is the speech-to-text service on Baidu's AI Open Platform, offered as APIs and SDKs for developers building Chinese-language voice features. It covers short voice commands, real-time streaming recognition and transcription of longer recorded audio files.

The service is split into several products: short speech recognition and a faster express version for brief utterances, real-time recognition for continuous audio, and audio file transcription for recordings. A self-training platform lets companies tune recognition with their own vocabulary, and far-field recognition targets smart speakers and other devices.

It sits next to Baidu's other speech tools, including speech synthesis, wake-word detection, speech translation and a call-center solution used for quality checks on phone calls. Access runs through a Baidu Cloud account in the console, and private on-premise deployment is offered for larger customers.