For those users encountering rate limiting on our DOC or Context 2.0 APIs or wishing to perform native language keyword search instead of the English machine translation searches of the API, we recommend using the Web NGrams 3.0 dataset.

Our APIs are rate limited to protect the underlying ElasticSearch clusters, given the enormous volume of requests we receive during peak events. In light of events in Ukraine, we've received a number of requests for guidance on performing at-scale querying, as well as from users wishing to perform native-language querying. Users needing to perform high volume querying should use the Web NGrams 3.0 dataset, which is available as a downloadable unigram JSON dataset covering all coverage we monitor in all 152 languages, updated every minute. Since these are downloadable, you can perform unlimited querying and since they are native language, you can search in the original source language.