Transcribe audio to text and generate expressive speech on-device via ominix-api.
npx skills add https://github.com/mofa-org/mofa-skills --skill voice-mofa-org
Please help me install this Agent Skill. Skill: voice Source: https://github.com/mofa-org/mofa-skills/tree/main/mofa-fm/.octos/platform-skills/voice Command: npx skills add https://github.com/mofa-org/mofa-skills --skill voice-mofa-org
On-device speech-to-text transcription and preset-voice text-to-speech with emotion control enable offline, private, low-latency voice workflows on Apple Silicon.
Transcribe an audio on-device with ASR or synthesize speech from text using a preset voice.
High-intent search queries and answers about installing and using this skill.
On-device speech-to-text transcription processes audio files or streams locally on Apple Silicon without network dependency. It uses the ASR endpoint via ominix-api to load models and transcribe audio entirely offline.
Offline text-to-speech synthesis supports emotion control using preset voices with OminiX/Qwen3 models. It generates natural-sounding speech locally on Apple Silicon via the TTS endpoint without requiring internet.
Local model management involves listing, downloading, loading, and unloading ASR and TTS models. The ominix-api handles discovery and download operations to switch between voice synthesis endpoints on-device.
On-device voice synthesis does not support custom voice cloning or profiles. It only uses preset voices for safety and simplicity, while voice cloning tasks are handled by mofa-fm.
Preset voice offline TTS is limited to predefined voice profiles and lacks custom voice cloning capabilities. It operates strictly on Apple Silicon for on-device processing without network dependency.
octos-orgTranscribe audio to text and synthesize speech on-device with local models.
naveencreationTranscribe audio files to text using the Qwen3-ASR-0.6B model.
hewliyangAutomate end-to-end voice cloning workflows on Apple Silicon with Qwen-tts.
guoqiaoServe OpenAI-compatible audio APIs for speech-to-text and text-to-speech on Apple Silicon.
