Hi everyone, I’m one of the maintainers of MyVoiceTyping, an open-source macOS voice typing app built around a local-first workflow.

Repo: https://github.com/botaruibo/MyVoiceTyping
Text polishing model: https://modelscope.cn/models/botaruibo/MyVoiceTyping-1.5b-q4
Fine-tuning dataset: https://github.com/botaruibo/MyVoiceTyping-Dataset

The app is focused on Chinese voice input for everyday writing: hold a hotkey, speak naturally, then let the app transcribe, restore punctuation, lightly correct/polish the text, and paste it back into the current input field.

The reason I’m sharing it here is that the current local ASR path uses the FunASR / SenseVoice ecosystem, with ONNX-based local inference in the packaged macOS app. For users who compare tools like Typless or other cloud voice-input products, the main difference is privacy: audio and text can stay on the Mac by default instead of being sent to a third-party cloud service.

What we are trying to optimize:

• local Chinese ASR for real writing / chat / notes scenarios
• punctuation restoration and lightweight text cleanup after speech recognition
• custom hot words for project names, technical terms, and personal vocabulary
• a fully open workflow where the app, model choice, and tuning dataset can be inspected or replaced

It is still an early open-source project, so I’d love feedback from SenseVoice/FunASR users, especially around local inference, packaging, latency, and Chinese dictation quality on different Macs.