Aratako/Irodori-TTS

Health v2 · maintenance only

Flags: no_releases

Adoption not part of the score

1219 stars · 151 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-30, confidence not recorded

Irodori-TTS is a Flow Matching-based text-to-speech model with training and inference code, built on a Rectified Flow Diffusion Transformer over DACVAE latents. It supports zero-shot voice cloning, multi-modal voice design with emoji-based style control, LoRA fine-tuning, and CLI/Gradio inference.

Use cases

• clone a voice from reference audio
• generate speech with emotion and style control
• design a custom voice from text descriptions
• fine-tune a TTS model with LoRA on my own voice
• train a flow matching text-to-speech model
• generate japanese speech from text
• run tts inference from the command line

When to choose

• you need zero-shot voice cloning or voice design with style/emotion control
• you want to fine-tune or train a modern flow-matching TTS model
• you want emoji-annotated text to influence speech delivery
• you need a compact open TTS model with released checkpoints and demos

When to avoid

• you need a production-ready OpenAI-compatible serving API out of the box (use the companion Irodori-TTS-Server)
• you need non-Japanese or multilingual TTS
• you have no GPU and need fast low-latency synthesis
• you just want a plug-and-play TTS without touching model code

Facets

library · maturity active

tts machine-learning deep-learning audio-processing llm-training speech-processing machine-learning python cli cross-platform flow-matching diffusion-transformer voice-cloning voice-design emoji-style-control dacvae lora-fine-tuning japanese-tts audio-watermarking gradio audio natural-language-processing gpu

Member repositories

For agents

markdown · JSON · MCP: product_card(name="Aratako/Irodori-TTS")