Soniqo Homebrew Tap

Homebrew formulas for Soniqo on-device speech tools.

Install

brew install soniqo/tap/speech

That installs:

• speech — CLI for ASR, TTS, speech-to-speech, voice cloning, VAD, speaker diarization, speech enhancement, forced alignment, wake-word detection, and translation
• speech-server — local HTTP / WebSocket server with an OpenAI Realtime API–compatible endpoint

Run speech --help or speech-server --help to see the full surface.

Requirements

• macOS 15 (Sequoia) or newer
• Apple Silicon (M-series chip)

Models are downloaded automatically from Hugging Face on first use and cached under ~/Library/Caches/qwen3-speech/.

Formulas

Formula	Description	License
speech	AI speech toolkit (Apache-2.0) — built from soniqo/speech-swift	Apache-2.0

Source

Library source lives at soniqo/speech-swift. The tap is auto-bumped on each upstream release.

License

The tap itself is Apache-2.0. Individual formulas may install software under their own licenses — see each formula's license line.