Back home
Security

Security & data handling

How PatentAgents.ai stores project data.

Data handling

PatentAgents.ai stores your source disclosures and generated drafts with your project.

Model training

Drafting runs on the Anthropic API, so your project content is sent to Anthropic for processing. Anthropic's Commercial Terms state that Anthropic may not train models on customer content from its services. Anthropic publishes its API data retention terms.

Retention

Archiving a project removes it from your project list; it does not delete the project's files, drafts, or drafting history. There is no automatic, time-based purge. To request deletion of your data, contact us at security@patentagents.ai.

Data residency

The web app is hosted on Vercel. Background services run in Google Cloud's us-east4 region (Northern Virginia, USA), and the database runs in Neon's AWS us-east-1 region (Northern Virginia, USA).

Access & separation

Each project is scoped to the account that owns it; project data is requested through authenticated, ownership-checked endpoints.

Account ownership

Workspace role labels (Owner / Attorney / Reviewer / Viewer) are currently display-only. Per-role permission enforcement is in progress and not yet active. Today, access is controlled by account ownership: only the account that owns a project can open it.

Administrative access

Access to customer project data is restricted to members of the small core team who hold explicit access to the production hosting project and its database credentials. There is no broader internal access; operational access is limited to support and debugging needs and is gated by hosting-project permissions and credentials.

Project history

Drafting is AI-assisted. Each step is saved on the project so you can read and revise the patent application.

Export controls

You can export the patent application as an editable Word .docx.

Exports are scoped to the authenticated account that owns the project. Export events are not currently recorded in an audit log.

Security contact

To report a security concern or request more detail on any of the above, reach out directly. We aim to acknowledge reports within one business day and provide a triage update within three business days.

security@patentagents.ai