Your data.
Your boundaries.

Customer data, backups, and logs stay within a defined hosting region. Hosting location is part of the workspace’s infrastructure configuration.

Connection credentials are encrypted using Cloud KMS. They are excluded from application logs and are never returned to AI assistants.

Data ingestion, storage, and AI access have separate responsibilities. Workspace checks apply wherever customer data is accessed.

Connections, datasets, exports, and audit events belong to a workspace. Access depends on current workspace membership and permissions.

Give each assistant access to specific datasets. Everything outside that scope remains inaccessible.

EverySchema enforces access permissions and query limits independently of the AI model. A prompt cannot grant additional permissions.

ChatGPT and Claude can query authorized data. They cannot change or delete records in your connected tools.

EverySchema connects your data to the assistant you already use. Your AI provider’s account settings and terms govern how that provider handles information returned to it.

Queries run with scoped access and limits. Credentials for your original data sources are not exposed to the assistant.

Review which datasets were queried and when through your workspace’s audit log.

Technical events support troubleshooting and traceability. Connection credentials are excluded from application logs.

Source, reporting period, currency, and time zone remain explicit. Compatible datasets can be combined without hiding differences in their definitions.

Removing a workspace includes removing its stored dataset rows. Deletion across storage systems and backup retention must be addressed together.

Retention and deletion requirements belong in the applicable data processing documents. Review these details for the data categories and services your team uses.