Features

We are constantly refining and expanding the range of our features offered. Please check back for more updates!

Large Language Model(LLM) based Features

Sentiment Analysis

a large language model is used to assess the sentiment of the transcribed interaction between the parties in the audio. the LLM is also asked to justify its sentiment analysis.

Summarisation

a large language model is tasked with generating a summary of the transcriptions generated from an uploaded audio file.

Speech-to-Text(STT) Features

ITN Formatting

inverse text normalisation(ITN) processing is used to format dates, times, decimals, currencies, emails and phone numbers in the transcribed text for improved readability.

PII Removal

model recognises personal identifiable information(PII) and removes them from the transcripts generated.

Speaker Diarization

model identifies and distinguishes between different speakers in the audio content. speaker tags for different speakers are included in transcripts generated.

Punctuation

model adds capitalisation and punctuation to sentence level outputs for better readability.

Confidence Scores

model generates word and sentence level confidence scores representing the level of certainty/confidence in the predicted text.

Multichannel Audio

process and transcribe multichannel audio files, generating transcripts for each channel with accurate timestamps.

Timestamping

model generates accurate word and sentence level timestamps for greater ease during review of transcripts generated.

Deployment Options

On-Prem Deployment

deployment of api and models in your on-prem servers; data stays secure within your firewalls.

Cloud based Deployment

deployment can be done on your preferred cloud platform vendor such as Microsoft Azure, Amazon AWS, etc

Language Packs

Singaporean English/Chinese

single model trained to simultaneously recognise both english and chinese language spoken by Singaporean accents.

Singaporean Chinese

model fine-tuned to recognise and transcribe chinese language spoken by Singaporean accents.

Bahasa Indonesia

model fine-tuned to recognise and transcribe bahasa indonesian language spoken by Indonesian accents.

Bahasa Melayu

model fine-tuned to recognise and transcribe bahasa melayu language spoken by Singapore/Malaysian accents.

Mainland Chinese

model fine-tuned to recognise and transcribe chinese language spoken by native speakers from mainland China.

Singaporean English

model fine-tuned to recognise and transcribe english/singlish language spoken by Singaporean accents.

Others

Transcript Editor

review and edit transcripts