Sakura AI Engine is a generative AI inference API from SAKURA internet that lets developers add AI features to their apps without setting up any of their own servers. The whole service runs in SAKURA internet’s Japanese data centers, so companies and local governments can keep confidential and personal data inside Japan.

A single API gives access to several foundation models, including domestic and open-source options, so teams can pick the model that fits each task instead of being tied to one. The same API covers text generation, audio transcription, text embeddings, and speech synthesis with VOICEVOX character voices. The endpoints follow the same conventions as OpenAI and Anthropic, which keeps the move from an existing setup simple, and customer data is never used to train the models.

The service can ground answers in a company’s own documents using Retrieval-Augmented Generation, linking a vector database to the models through the API. This suits internal chatbots and FAQ tools that need to draw on private company knowledge.

A free plan covers a set amount of usage each month at no cost, which is handy for small projects and for trying the service before committing. Signups for the free plan are subject to a capacity cap. A browser-based Playground also lets teams test models directly before building anything.