Blog

Guides for private, local AI voice work.

Voice Cloning

How to Run Chatterbox TTS Locally on Mac

A practical guide to running Chatterbox Turbo, the original Chatterbox model, and Chatterbox Multilingual locally on a Mac with Python, Apple Silicon MPS acceleration, and CPU fallback.

Jun 1, 202612 min read
Voice Cloning

How to Run CosyVoice 3 Locally on Mac

Learn how to run CosyVoice 3 locally on a Mac with speech-swift, MLX, mlx-audio-plus, Rust Candle bindings, and the official FunAudioLLM Python repository.

Jun 1, 202614 min read
Voice Cloning

How to Run Fish Audio S2 Pro Locally on Mac

Learn the practical ways to run Fish Audio S2 Pro locally on a Mac, including MLX on Apple Silicon, native Swift integration, experimental GGUF and Metal ports, and the official CUDA self-hosting route.

Jun 1, 202614 min read
Local TTS

How to Run Kokoro TTS Locally on Mac

Learn the practical ways to run Kokoro TTS locally on Mac, including Python with PyTorch, ONNX Runtime, MLX on Apple Silicon, JavaScript, native Swift integrations, and ready-made desktop apps.

Jun 1, 202614 min read
Voice Cloning

How to Run Qwen3-TTS Locally on Mac

Learn how to run Qwen3-TTS locally on a Mac with MLX, the official Python package, the local Web UI, and native Swift integration. Compare model sizes for voice cloning, preset voices, and description-based voice design.

Jun 1, 202614 min read
Local TTS

How to Run Orpheus TTS Locally on Mac

Learn how to run Orpheus TTS locally on a Mac with orpheus-cpp, Metal acceleration, LM Studio, llama.cpp, GGUF models, community Web UIs, and the official Python package.

Jun 1, 202613 min read
Models

CosyVoice 3: Multilingual Zero-Shot TTS

A deep technical exploration of Alibaba's CosyVoice 3: the supervised multi-task MinMo+FSQ speech tokenizer at 25 Hz, the three-stage LLM → DiT flow matching → HiFi-GAN pipeline, DiffRO differentiable reward optimization, 1M-hour data scaling across 9 languages and 18+ Chinese dialects, and zero-shot voice cloning.

May 30, 202614 min read
Models

Fish Audio S2 Pro: Benchmark-Leading Voice Cloning

A deep technical exploration of Fish Audio S2 Pro: the dual-autoregressive master-slave architecture, multi-stage training (semantic pre-training, speech captioning, GRPO alignment), the 15,000+ natural language tag system for prosody control, in-context voice cloning, and benchmark results that top the TTS Arena leaderboard.

May 30, 202612 min read
Local TTS

TTS for Elderly Users: Supporting Aging in Place with Local Voiceover on Mac

How text-to-speech helps older adults stay independent at home — reading medication instructions, personal correspondence, news, and books aloud with privacy-safe offline TTS on Mac.

May 27, 20267 min read
Voice Cloning

The Complete Guide to Text-to-Speech (2026)

A complete guide to text-to-speech in 2026 — how TTS works, types of TTS systems, common use cases, open-source model rankings, cloud API rankings, local vs cloud tradeoffs, how to choose the right solution, and where TTS is headed.

May 26, 202618 min read
Guides

Common Use Cases for Text-to-Speech in 2026

A practical overview of the most common text-to-speech use cases in 2026 — accessibility and reading support, narration and voiceover, proofreading and review, and private professional workflows.

May 26, 20268 min read
Local TTS

Does Spokio Work on M1, M2, M3, M4, M5, and Intel Macs?

Check whether Spokio works on your Mac, including Apple Silicon and Intel models, and learn the macOS requirement for offline text-to-speech.

May 25, 20264 min read