#语音输入法以后 Spokenly 就是小狼毫了。不过非 macOS 键盘如何映射 Fn 键啊，键盘固件里的 “切换 Fn 层”按键，并不对应原生 Fn 键，只是有切换 Fn 的功能...产品经理快给自定义选项... https://t.co/bLYN2Vkrwx𝘁𝗮𝗿𝗲𝘀𝗸𝘆 (@taresky):#AI某种程度上，AI 语音输入法的竞争已经结束了，还没卖产品拿融资的应该尽快。有两种模式：1. Spokenly 这类全功能客户端（免费用户自接第三方 API 或者本地模型，付费用户可免配置体验良好）现在每个月，最好的语音识别模型都在变化，这是唯一合理的形态。 ...

RT 𝘁𝗮𝗿𝗲𝘀𝗸𝘆 (@taresky)Soniox v4 的模型发布了。可惜这个模型并不是一个实时转录的模型，看起来它的优点是对上下文识别会更加准确。我测试了自己常用的语音输入内容，不管是中文、英文、生僻词汇、Crypto 领域的专业词汇，整个识别结果乃至标点符号都没有出现错误。当之无愧的当前最强语音识别模型。 且免费。以上内容通过该模型识别并输入。Spokenly 的配置方法如图2。 ...

Soniox v4 的模型发布了。可惜这个模型并不是一个实时转录的模型，看起来它的优点是对上下文识别会更加准确。我测试了自己常用的语音输入内容，不管是中文、英文、生僻词汇、Crypto 领域的专业词汇，整个识别结果乃至标点符号都没有出现错误。当之无愧的当前最强语音识别模型。 且免费。以上内容通过该模型识别并输入。Spokenly 的配置方法如图2。 ...

Core Viewpoint - Alibaba's Qwen has officially open-sourced the Qwen3-ASR series of models, which are powerful speech recognition models designed to support multiple languages and dialects [1] Group 1: Model Details - The Qwen3-ASR series includes two comprehensive speech recognition models: Qwen3-ASR-1.7B and Qwen3-ASR-0.6B [1] - Additionally, the series features an innovative speech forced alignment model, Qwen3-ForcedAligner-0.6B [1] Group 2: Language Support - The Qwen3-ASR models support speech recognition and language identification for 52 languages and dialects [1]

Core Viewpoint - Rongke Technology (300290) is focusing on the development and operation of nursing service robots, which are designed to enhance medical services through intelligent automation [1] Company Overview - The company's nursing service robots are developed and operated by its wholly-owned subsidiary, Shenzhou Shihan [1] - The robots are capable of performing various tasks such as ward education, patrol, item delivery, and bedside interaction [1] Key Features - Core functionalities of the robots include interactive Q&A, intelligent reminders, patient guidance, video-based visiting services, and data recording and analysis [1] - The robots utilize the DeepSeek local knowledge base, Shihan's self-developed voice recognition model, and a multimodal AI vision model to achieve intelligent medical services [1]