AI Disclosure

Everything is synthetic

Every video produced here is generated by a machine-learning model from a text prompt. No footage is filmed. Videos may look realistic; they are not recordings of real events.

Which models

We route generations to third-party providers through fal.ai (Kling by Kuaishou, Veo by Google DeepMind, Seedance by ByteDance) and Replicate. The model used is shown on every generation. Providers change; the current catalogue is on the Models page.

Labelling

Videos we publish on social media are labelled as AI-generated using each platform's disclosure tools, as required by their policies and by the EU AI Act (Article 50). Some providers embed invisible watermarks (e.g. Google SynthID) or C2PA metadata that identify the output as synthetic; these cannot be removed.

Screening

Prompts are screened by an automated classifier (Google Gemini) before any credits are spent, and again by the provider's own safety filter. We block sexual content, sexualised minors, real identifiable people, brands and logos, hate, graphic violence and illegal instructions. Screening is imperfect; report problems to support@kinomuse.com.

Copyright

Under current US Copyright Office guidance, purely prompt-generated output is not protected by copyright. Your creative selection and editing may be. We make no representation about the copyright status of any output in your jurisdiction.