Release 114 - October 1, 2026

What's new in this release:

Support of Soniox speech-to-text 

Wildmoka offers its own speech-to-text engine or can support Deepgram, Speechmatix, and Rev.ai. Wildmoka now adds the support of Soniox as a speech-to-text provider for live streams and recorded content. Soniox uses a single multilingual model covering more than 60 languages and can recognize multiple languages within the same conversation—without switching models.

How it works: Soniox is a third-party service. Customers need their own Soniox subscription and API credentials, same as for other third-party speech-to-text providers. Contact your Backlight representative to configure Soniox for your Wildmoka environment. 
Once a Soniox API key has been configured for your Wildmoka environment, Soniox can be selected as the speech-to-text provider for a stream or file workflow. Transcriptions are then available in the new Clip Editor, including:

• Captions displayed on the video preview
• Transcribed words in the Captions tab
• The existing Wildmoka workflows powered by speech-to-text

Why this matters: Soniox gives multilingual teams another transcription option, with broad language coverage and support for conversations that naturally move between languages. It can be particularly useful when existing providers do not fully address a language, region, or multilingual use case.

Feature available to all customers from October 1, 2026, subject to Soniox configuration and subscription.

Improvements and fixes

This release makes the Clip Editor more stable, with a smoother timeline and preview, no more black thumbnails or flashes when seeking, and clearer error handling that separates network issues from expired media. Publishing is more reliable, with metadata correctly populated from the selected template and improved thumbnail generation and publication. In Channel Studio, continuity playlists are improved and stream playback no longer pauses, slows down or stops.