moss-tts

Audiobook creation app supporting too many TTS models (Qwen3-TTS, OmniVoice, VibeVoice, etc), focused on high-quality output. Plus audio-synced reader web app and standalone server component.

ComfyUI nodes for MOSS-TTS-Local-Transformer-v1.5 , a 48 kHz stereo local Transformer TTS model with direct generation, zero-shot voice cloning, continuation, duration control, pronunciation markup, multilingual synthesis, and code-switching.

CPU-only inference pipeline for MOSS-TTS 8.4B in PyTorch, forcing Float32 to bypass BFloat16 issues and adding selective quantization on the LM backbone. Runs efficient text-to-speech on hardware without a GPU.

Roo Voice — local, reference-conditioned single-voice TTS. One-click app + models for Apple MLX, NVIDIA CUDA, AMD Vulkan/ROCm, and NPU (ONNX).

The best voice-cloning TTS on the planet. Clone any voice, 31 languages, runs 100% offline. Free forever.

🔊 Local cloned-voice feedback for Claude Code + Codex on Apple Silicon — MOSS-TTS-Nano on MLX, 0.3s first sound, spoken summaries, multi-session queue, offline after download.

Auditable script-to-video workflow with MOSS-TTS, image keyframes and HyperFrames