twenty/v2.41.0 v6.2.2
snakers4/silero-vadSep 17, 2026 → Sep 17, 20260 days between releases
What's New
Adds offline ONNX sequence model with 16 kHz support and new API function
Breaking Changes
- load_silero_vad now requires sequence=True parameter to load the new model
- Model file is now separate: silero_vad_16k_sequence.onnx
New Features
- Offline ONNX sequence model for speech timestamps
- 16 kHz audio support for sequence processing
- New get_speech_timestamps_sequence function
- Dedicated model file for sequence processing
Upgrade Notes
Update code to pass sequence=True to load_silero_vad and ensure audio is sampled at 16 kHz. The new model file must be downloaded separately.
The Two Releases
twenty/v2.41.0
Sep 17, 2026· 1 month ago
A major update to the Twenty platform featuring workflow validation, usage limit management, and extensive messaging/AI enhancements, alongside a breaking change regarding workspace connection management.
Full release notesv6.2.2: Sequence silero-vad
Sep 17, 2026· 1 month ago
This release introduces a new offline ONNX sequence model optimized for 16 kHz audio processing. It includes a dedicated model file and a new API function to retrieve speech timestamps efficiently.
Full release notes