⬢github Python · 56.4K ★ +1.1K since we first saw it · pushed 29 d ago · MIT
microsoft/VibeVoice
Open-Source Frontier Voice AI
VibeVoice is Microsoft's open-source voice AI suite, best known for text-to-speech and speech recognition. It includes VibeVoice-ASR, a speech-to-text model that transcribes up to 60-minute audio in one pass with speaker labels, timestamps, 50+ languages, streaming recognition, hotwords, and finetuning, plus a BitNet-compressed CPU-only edge variant.
Why now: Recent releases like VibeVoice-ASR-Streaming and the CPU-only BitNet engine, plus integrations into Hugging Face Transformers and Azure AI Foundry Labs, are drawing attention.
Who it is for: Developers building speech recognition, transcription, or TTS features, especially those needing long-form or on-device audio processing.
Stars over our 38 snapshots: 55.4K to 56.4K, since 1 d ago.
Where people talked about it
- ⬢github microsoft/VibeVoice 12 h ago
API: https://socialmediatrends-api.osmike.com/v1/repos/microsoft/VibeVoice