Changelog

The latest changes to the CCV AI Services

2025-10-24 - Transcribe v1.0.3

This update consolidates Transcribe's multi-language support for both models based on user feedback and usage data. These languages include:

• Arabic

• Korean (for Whisper model only. Gemini does not support Korean at the moment)

• Vietnamese

We have also removed language options that are not supported by Gemini, including Cantonese and many others.

Accessibility was also improved in this update and multiple minor UI issues are fixed

2025-10-17 - Transcribe v1.0.2

This update fixes the issue where transcription with the Whisper model could fail for some languages. The alignment model for these languages are now hosted locally on Google Cloud. Whisper now reliably supports these languages:

• English

• Spanish

• French

• German

• Italian

• Japanese

• Russian

• Dutch

• Portuguese

• Chinese (Mandarin)

The users can now request for more languages if the language they want to transcribe is not listed above.

There are other small UI tweaks:

• The "View Job" and "View Transcription" page now displays the time used to process the job/transcription and the associated Real Time Factor (RTF).

• Carbon emission information is not displayed on the "View Job" page until the transcription job is completed for accuracy.

• The FAQ section is not displayed on smaller screen sizes to save screen space.

• Other small language fixes.

2025-10-08 - Transcribe v1.0.1

This is mainly a maintenance update that changes some of the languages around the use of the service and adds a 20-hour usage limit to prevent the abuse of the system.

2025-10-01 - AI Transcribe Service goes 1.0

Over 17 months in development, our Transcribe Service is officially released with version 1.0! This official release version is packed with new features frequently requested by our users!

• You can now edit your transcription while listening to the audio (when available)! You can add, remove, and name speakers, insert/edit segments, and edit timestamps!

• You can play your audio files now. The sentence being played will be highlighted, and transcriptions can even be automatically scrolled so that the segments/sentences are always centered on the screen.

• Azure Speech-to-text model is removed from the service. According to our experience, the battle-tested Open Whisper model is significantly faster and more accurate. We removed the Azure model to simplify the model selection process.

• Gemini transcriptions might be even faster now.

• We also fixed an error where consecutive OpenAI Whisper jobs submitted by one user may fail.

Thank you for using CCV AI Services! We will continue to improve with the feedback from all of you!