Synthetic Media & Deepfake Policy

Nastia is an adult service for users aged 18 and over. All companions and media on the Service are AI-generated and depict fictional adults. Using the Service to create non-consensual or deceptive synthetic media of a real person is strictly prohibited.

Prohibited Uses

You may not use Nastia to do any of the following:

  • create sexual or intimate imagery of a real, identifiable person without that person's consent (non-consensual intimate imagery, or "NCII");
  • upload or supply a photograph of a real person in order to reproduce that person's likeness, including in sexual or suggestive content;
  • impersonate, defame, or deceptively depict any real person, or create any media intended to be mistaken for a real person (a "deepfake");
  • infringe any person's privacy, publicity, or likeness rights; or
  • create any depiction that sexualizes a minor, including AI-generated depictions, which is covered by our zero-tolerance Child Safety Policy.

Controls We Apply

Nastia is designed to create original AI characters, not to reproduce real people. We operate the following controls at the point of generation, independently of any report:

  • Face references are AI-generated only. A face reference can only be selected from images that Nastia itself generated for the user's companion. The Service does not accept user-imported, uploaded, or external photographs as face references, so a photograph of a real person cannot be used to guide a companion's appearance. Because every accepted face reference originates from Nastia's own generation pipeline, it depicts a fictional AI character and not a real person.
  • Pose references are de-identified. When a user provides a pose reference, the system reduces it to an abstract pose skeleton (geometric body, hand, and facial keypoint positions only, with no facial texture, color, or other identity-bearing detail) and replaces the stored reference with a heavily blurred, low-resolution version. We retain only this de-identified reference plus the skeleton, never a faithful copy of the original, and it is the skeleton (not a face or identity) that guides generation.
  • No public exposure. The Service is private to each user: uploaded reference images are never shown in any public or shared gallery, and there is no public feed or sharing of content between users.
  • Real people and public figures are blocked. Prompts that name a real, identifiable person or public figure are automatically refused at the point of generation, and generated media is reviewed for resemblance to known public figures. Nastia is built to produce original fictional characters, not to depict real individuals.
  • Moderation. Automated and manual moderation operate at the point of generation to block prohibited content, and we continuously refine these safeguards.

How a pose reference is de-identified

A user-provided pose reference next to what Nastia retains: a heavily blurred, de-identified image overlaid with only an abstract pose skeleton.
A provided pose reference is used transiently to extract an abstract pose skeleton. Nastia stores only a heavily blurred, de-identified reference plus the skeleton, so no identifiable likeness is retained.

Reporting and Enforcement

Where we become aware of content that appears to violate this policy, we:

  • remove the content;
  • suspend or terminate the responsible account;
  • preserve relevant records where appropriate; and
  • cooperate with law enforcement and respond to valid legal process.

Reports that a real person has been depicted without consent, including alleged NCII or deepfakes, are prioritized. We aim to action valid take-down requests promptly, and within five business days, consistent with our Complaint Policy.

How to Report

If you believe content on Nastia depicts you or another real person without consent, or otherwise violates this policy, contact us at [email protected] or use our Content Take-Down Policy. Please include enough detail for us to locate the content.

Related Policies

See our Child Safety Policy, Content Take-Down Policy, Complaint Policy, and 2257 Compliance Statement.