Public preview
The planned anonymous snapshot reads a bounded set of public web pages. It does not accept credentials or access private content. Raw preview crawl content is intended to expire after seven days, with aggregate abuse and cost counters retained separately.
Connected accounts
Shopify, GitHub, Search Console, analytics, and deployment access will use OAuth or platform-native installation flows. Tokens must be encrypted, scoped to the requested job, and revocable.
The initial Shopify product does not require customer or order data. Write permissions are requested progressively only when execution is enabled.
Tenant isolation and audit history
Organizations, sites, evidence, approvals, executions, and usage remain tenant-scoped. Material actions produce append-only audit events. Product authorization is enforced server-side rather than by hiding interface controls.
Model and research providers
Crawled content is treated as untrusted input. Provider use, retention, and training controls will be documented before the live beta. Customers receive Foliora's derived research and work—not a replicated third-party data product.
Contact
Privacy and deletion questions can be sent to privacy@foliora.ai. This draft is not a substitute for the final legally reviewed privacy policy.