Know what happens before you import a client manuscript
Editors need a concrete processing boundary, not a general promise that AI is safe. This page summarizes the current product behavior and links to the controlling privacy and terms pages.
Import and storage
Supported manuscript files are parsed in the browser so the user can review the proposed chapter split before confirming import. Confirmed manuscript content and project data are stored so Novilot can provide the workspace, analysis, and later access described by the product.
The current product is not a local-only application. Do not represent it to a client as equivalent to an offline Word add-in.
AI processing
Relevant manuscript content is sent to configured AI providers when the user deliberately runs an AI-backed analysis. Novilot’s privacy page states that manuscript content is not used to train AI models. Editors should still disclose the provider-assisted workflow when their contract or professional obligations require it.
Analytics boundary
Novilot’s product analytics are configured to avoid manuscript prose: automatic text capture is disabled, session-replay text and inputs are masked, and manually captured events are limited to non-prose properties such as counts, IDs, booleans, and enums.
Deletion and editor responsibility
Projects remain available until the user deletes them or the account lifecycle described by the product applies. Editors should delete client projects when the engagement and retention obligation end, maintain their own authorization record, and preserve the original client file outside Novilot as appropriate.
Continue the editor workflow
Use only a client-authorized manuscript
Confirm that the current processing boundary fits the engagement before starting an editor preflight.
Run a free manuscript preflight