Tested ZONOS2 on an 8GB RTX 4060 Laptop (WSL2): the 15.3GB bf16 weights run via Windows system-memory fallback, a KV-cache override, and the CUDA toolkit at ~1/20 realtime. Plus a Japanese name-accent gotcha with A/B audio.
VoxCPM2 sits in the tokenizer-free corner. Mapped vs F5-TTS, CosyVoice2, Irodori-TTS, Style-Bert-VITS2; plus why Japanese TTS still leans on OpenJTalk.
NII/LLMC released CC Audio and Archive.org Audio Dataset. URL lists, metadata, and a downloader covering 48,000+ hours of Japanese audio. What it actually contains and how it fits into TTS, ASR, and audio model training.
An open-source TTS model distilled from the ZipVoice architecture into four inference steps, delivering voice cloning with 1 GB of VRAM and 150x real-time speed. It also compares itself with the other TTS models covered on this blog.
A local-first voice cloning, TTS, and audiobook app that brings Qwen3-TTS, Chatterbox, Kokoro, and IndexTTS-2 into a single GUI. It uses a FastAPI backend, Flutter UI, and an MCP server.
MioTTS from Aratako is a family of 0.1B to 2.6B Japanese-English TTS models built from scratch around the custom MioCodec. Its key feature is that it runs directly in llama.cpp and Ollama.
A technical overview of Qwen3‑TTS from Alibaba’s Qwen team: one‑line pip install, 3‑second voice cloning, natural‑language voice design, and support for 10 languages including Japanese. Apache 2.0 licensed.
Overview of PersonaPlex‑7B‑v1 released by NVIDIA in January 2026. A Moshi‑based voice dialog model that enables full‑duplex conversation and persona control.
The Web Speech API + Gemini + VOICEVOX setup is complete — an AI character you can actually have a voice conversation with. Key implementation notes and impressions.
