Aims to reduce the persistent storage of model call content on the platform. Any user can apply for activation…
By Sam Ovie
·
Published on September 20, 2026
·
2 min read
Zhipu MaaS platform announced today that it will soon launch a data content non-retention function to provide stricter data privacy protection for enterprise and developer users.
The announcement states that the data content non-retention mechanism aims to reduce the persistent storage of model call content on the platform. Any user can apply for activation. Once activated, the Zhipu MaaS platform will not statically store user inputs and outputs; the data will only be used to complete the current model call. Enterprises and developers can submit activation applications through the MaaS console. The specific effective time and applicable scope are subject to the platform’s confirmation.
The announcement states that this feature does not mean that data will not be retained under any circumstances. Functions such as Batch API and File API that require persistent storage of tasks or files on the platform side are not covered above. In cases where retention is required by laws and regulations, or to verify suspected violations or abuses, the platform may retain relevant data for 30 days or more as required.
As previously reported, ZCode, a programming product under Zhipu, apologized to affected users on September 18 through the official Zhipu group in response to discussions in the community regarding code repository data uploads. The company also issued a statement saying that it had completed a self-inspection immediately and the issue had been fixed.
The problem stemmed from ZCode’s “Code Repository Index” feature, which aims to help users generate repository indexes locally to support features such as session checkpoint recovery (including historical versions), historical version rollback, and Repo Wiki (code repository knowledge base).
