Kugel-2 VibeVoice model pulled from Hugging Face

The sudden removal of Kugel-2 highlights the fragile nature of community-driven model releases and the tension between open access and monetization.

• –High-fidelity cloning with minimal audio makes it a top-tier choice for local TTS enthusiasts
• –Built on Qwen2.5, demonstrating the versatility of LLM backbones for specialized audio tasks
• –Integration with ComfyUI's TTS-Audio-Suite has already cemented its place in local creative workflows
• –The "mistaken upload" narrative often masks licensing or commercial pivot conflicts in the open-source space
• –Community mirrors are already proliferating, making the takedown largely symbolic but technically annoying