The Speech Lab is developed in collaboration with Professor Li Haizhou (National University of Singapore) and Associate Professor Chng Eng Siong (Nanyang Technological University).

Speech recognition helps you convert audio to text. Our speech engine recognises and transcribes English, Mandarin and Singlish. Our code-switching technology is uniquely catered for Singlish as it can recognise conversations comprising words from different languages (a mix of English and Mandarin). This speech engine can be further customized for various use cases in different domains.

Tip: Our speech engine performs best in quiet environments. For optimal results, users should use a high-quality microphone and be as conversational as possible when recording.

Use Cases

Automatic Transcriptions

Speech-to-text at Call Centres
Transcribing during interviews
Medical Consultations

Chatbot And Digital Assistants

Transcribing voice commands in chatbots and smart assistant technology

How it Works

Batch Transcribe

Analyses completed audio recording for transcription.

Transcribes audio data in batches
High-score transcriptions
Supports .wav, .mp3, .mp4 files

Live Transcribe

Analyses audio stream in real-time to provide on-the-fly transcription.

Real-time transcription
Supports .wav files

Online Demo!

AI SIngapore and Speech Lab team have put up an online demonstration of the SInglish speech engine to allow interested users to test the system. Please note that this system may not run the latest engine, and may not be tunned for your specific domain. For higher accuracy, the Speech Lab team can undertake a project under the 100E programme to optimize the engine to suit your specific domain and requirements.

Create an account or log in using your Gmail account.