Transcribe – The Speech-to-Text service for Whole-of-Government | Singapore Government Developer Portal
Have feedback? Please

Transcribe header banner

Transcribe is a Speech-to-Text (STT) platform with auto-transcription technologies and localised STT services across various speech models and data classification. Its Application Programming Interfaces (APIs) enable integration with other systems.

Transcribe can be used to produce transcripts of interviews, speeches and meeting minutes. Given the confidential nature of government documents, the platform is designed to be secure, yet easily accessible.

Key Features

Transcribe aims to improve the productivity of public officers with these unique offerings:

  • Available on both Internet and the Intranet environment for flexible use
  • Hosts up to RESTRICTED and SENSITIVE (NORMAL) data
  • Auto-onboard all public officers via WOG AAD authentication

Other key features include:

  • Open to addressing agencies’ needs through collaboration on operational use cases and model adaptation
  • Live and media file transcription
  • Workspace collaboration across Whole-of-Government
  • Multi-language support

Success Story

Transcribe was first used in Parliament Sittings to generate verbatim transcripts for Parliament users to produce the Hansard Report. This was done by integrating Transcribe into the workflow. By using Transcribe, Parliament users were also able to reduce the man hours from 2.5hrs to 1.5hrs for each 15 min transcript produced.

In its pilot phase, Transcribe also received positive feedback and a 75% satisfaction rate from a pool of 100 users across 10 government agencies.

Transcribe achieved more than a 200% year-on-year growth when it comes to the number of registered users. At the end of Q1 in FY22, close to 6,000 tasks were completed on transcribe, which amounted to more than 1,600 hours of transcription length. As of Q2 in FY22, Transcribe has a total of 1,032 registered users across 72 agencies.

Users can look forward to new transcript editing and speaker diarisation functions by end of 2022.

Pricing

Transcribe is free to use for now. A pricing strategy may be implemented in late 2024.

Contact Us

Reach out to the product team with your queries or feedback through this form.


Last updated 26 January 2024


Was this article useful?
Send this page via email
Share on Facebook
Share on Linkedin
Tweet this page