Data retention policy
We retain Customer Information, Customer Data and the communications and analytics generated by use of our Service (collectively, the “Service Output”), as long as we believe it is necessary and relevant for the operation of our Service. Service Output is retained at a company and employee level. A Customer Administrator may request that Customer Information and/or Customer Data be deleted from our systems after the termination of a Customer’s subscription. (However, Customers should understand that Users may elect to export and retain their own feedback data.) Under certain circumstances we may retain Customer Information from terminated subscriptions. We do this to ensure that the Service is not interrupted and none of the Service Output is lost due to interrupted Service subscriptions or other unanticipated events. We may also retain Customer Information after a Customer subscription term ends to prevent fraud, collect any fees owed, resolve disputes, troubleshoot problems, assist with any investigation, enforce our Terms of Service, comply with applicable legal data retention obligations and take other actions permitted by law.
Data archiving and removal policy
We retain Customer Information, Customer Data and the communications and analytics generated by use of our Service (collectively, the “Service Output”), as long as we believe it is necessary and relevant for the operation of our Service. Service Output is retained at a company and employee level. A Customer Administrator may request that Customer Information and/or Customer Data be deleted from our systems after the termination of a Customer’s subscription. (However, Customers should understand that Users may elect to export and retain their own feedback data.) Under certain circumstances we may retain Customer Information from terminated subscriptions. We do this to ensure that the Service is not interrupted and none of the Service Output is lost due to interrupted Service subscriptions or other unanticipated events. We may also retain Customer Information after a Customer subscription term ends to prevent fraud, collect any fees owed, resolve disputes, troubleshoot problems, assist with any investigation, enforce our Terms of Service, comply with applicable legal data retention obligations and take other actions permitted by law.
Data storage policy
We retain Customer Information, Customer Data and the communications and analytics generated by use of our Service (collectively, the “Service Output”), as long as we believe it is necessary and relevant for the operation of our Service. Service Output is retained at a company and employee level. A Customer Administrator may request that Customer Information and/or Customer Data be deleted from our systems after the termination of a Customer’s subscription. (However, Customers should understand that Users may elect to export and retain their own feedback data.) Under certain circumstances we may retain Customer Information from terminated subscriptions. We do this to ensure that the Service is not interrupted and none of the Service Output is lost due to interrupted Service subscriptions or other unanticipated events. We may also retain Customer Information after a Customer subscription term ends to prevent fraud, collect any fees owed, resolve disputes, troubleshoot problems, assist with any investigation, enforce our Terms of Service, comply with applicable legal data retention obligations and take other actions permitted by law.
Data center location(s)
Republic of Korea
Data hosting details
Cloud Storage
Data hosting company
Amazon AWS
App/service has sub-processors
no
App/service uses large language models (LLM)
yes
LLM model(s) used
GPT-4, GPT-4o-mini
LLM retention settings
All data are stored in AWS Asia Pacific (Seoul) region.
LLM data tenancy policy
Customer data is logically isolated and secured. Interactions with model only involve bare minimum information used for user identification.
LLM data residency policy
Prompts are embedded into our API statically. Generated content is stored strictly for logging and monitoring purposes. No prompts or completions are used to train the model.