rinna is an AI company that develops and provides Japanese-optimized large language models and conversational AI services. Originally inheriting Microsoft’s chatbot AI business, rinna now operates independently to create models specialized for Japanese language and cultural understanding.

rinna provides commercially usable Japanese language models released under MIT License that teams can self-host and customize without vendor lock-in or licensing fees. These include the Llama 3 Youko series and Japanese GPT models published on Hugging Face.

For enterprise deployments, rinna provides commercial APIs through Tamashiru Enterprise for dialogue, sentiment analysis, and text classification, plus Koemotion for emotionally expressive text-to-speech. Custom LLM development services are available for domain-specific needs.