Algolia uses a modified version of Kuromoji thats been tailored for our needs. We do not expose Kuromoji APIs directly.
You can configure your Japanese search using the following settings:
1. To achieve tokenization at query time, you can set the queryLanguages parameter to [“ja”]
2. To achieve tokenization at indexing time, you can set the indexLanguages parameter to [“ja”]. This setting is not yet in our documentation, but you can use it in the setSettings method.
These settings will ensure Japanese word segmentation is achieved correctly, but will also allow you to benefit from Algolia's typo-tolerance and proximity features.