Getting Started

speech-swift provides on-device AI speech processing for macOS and iOS on Apple Silicon. Models run locally using MLX (Metal GPU) and CoreML (Neural Engine).

Requirements

• macOS 15+ (Sequoia or later)
• Apple Silicon (M1, M2, M3, M4 series)
• Xcode 16+ / Swift 6.0+
• 8 GB RAM minimum (16 GB recommended for larger models)

Installation

Homebrew (CLI)

The fastest way to try speech-swift on macOS. Installs both the speech CLI and the speech-server HTTP/WebSocket server (OpenAI-compatible /v1/realtime endpoint). Requires native ARM Homebrew (/opt/homebrew).

brew install speech

After install, both binaries are on your PATH:

speech transcribe recording.wav
speech speak "Hello, world!" --output hello.wav
speech-server --port 8080            # local HTTP / WebSocket server

Swift Package Manager

Add speech-swift to your Package.swift dependencies:

dependencies: [
    .package(url: "https://github.com/soniqo/speech-swift", branch: "main")
]

Then add the modules you need to your target:

.target(
    name: "MyApp",
    dependencies: [
        .product(name: "Qwen3ASR", package: "speech-swift"),
        .product(name: "Qwen3TTS", package: "speech-swift"),
        .product(name: "SpeechVAD", package: "speech-swift"),
        // ... add any modules you need
    ]
)

Available Modules

Building from Source

Clone the repository and build:

git clone https://github.com/soniqo/speech-swift.git
cd speech-swift
make build

Quick Start: Transcribe Audio

CLI

# Transcribe a WAV file
.build/release/speech transcribe recording.wav

Swift API

import Qwen3ASR

let model = try await Qwen3ASRModel.fromPretrained()
// audioSamples: [Float] PCM at 16 kHz (e.g. decoded from a WAV)
let text = model.transcribe(audio: audioSamples, sampleRate: 16000)
print(text)

Models are downloaded automatically from HuggingFace on first use and cached in ~/Library/Caches/qwen3-speech/.

Quick Start: Text-to-Speech

CLI

# Generate speech
.build/release/speech speak "Hello, world!" --output hello.wav

Swift API

import Qwen3TTS
import AudioCommon

let model = try await Qwen3TTSModel.fromPretrained()
let audio = model.synthesize(text: "Hello, world!", language: "english")
try WAVWriter.write(samples: audio, sampleRate: 24000, to: URL(filePath: "hello.wav"))

Model Downloads

All models are downloaded from HuggingFace on first use. Approximate sizes:

Next Steps

• CLI Reference — all available commands and options
• Qwen3-ASR Guide — detailed speech-to-text documentation
• Qwen3-TTS Guide — detailed text-to-speech documentation
• API & Protocols — shared protocols and types