Hugging Face releases open-source speech and decision models for Apple platforms

Hugging Face has released two new open-source Swift packages, AnySpeechRecognizer and AnyDecisionModel, designed for Apple platforms. These packages allow developers to integrate speech recognition and decision-making capabilities into their applications using various on-device models. AnySpeechRecognizer supports Apple's new speech-to-text API and offers alternative engines like NVIDIA's Parakeet and Alibaba's Qwen3-ASR, with features like speaker diarization. AnyDecisionModel handles yes/no, multiple-choice, and rating-scale questions, leveraging models that run on MLX, Core AI, or llama.cpp. AI

IMPACT Enables developers on Apple platforms to more easily integrate advanced speech recognition and decision-making capabilities using open-source models.

RANK_REASON This is a product release from a well-known AI platform, but it focuses on integrating existing models into a specific ecosystem (Apple platforms) rather than a novel frontier model release or significant industry-wide research.

Read on Hugging Face Blog →

• Alibaba Group
• AnyDecisionModel
• AnyLanguageModel
• AnySpeechRecognizer
• Apple Inc.
• Core MLL complex [nucleoplasm]
• Hugging Face Hub
• Mlx
• Nvidia
• Ollama
• parakeet
• Qwen3 ASR

• product
• model release

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →