AI Disclosure
Last updated: June 8, 2026
This page explains how Yotzer uses artificial intelligence, how AI-generated content is marked, and your rights and responsibilities. It is published in line with Article 50 of the EU AI Act.
What Yotzer is
Yotzer is an AI content production platform. It uses large language models and image generation models to help you create text, images, slides, and video in your brand voice.
Yotzer is operated by Michelle Sartorio, based in Israel, serving users in the EU and internationally.
AI models we use
Yotzer tools roll out in stages. As each tool launches it uses the providers below. Until a tool is live, no content is sent to its provider.
Text generation
Anthropic Claude (Claude Sonnet and Claude Haiku). Prompts and content are sent to Anthropic's API for processing.
Image generation
OpenAI GPT Image 2 (gpt-image-2). Prompts are sent to OpenAI's API.
Voice synthesis
Fish Audio, for video voiceover. Text is sent to Fish Audio's API.
In-app assistant (Zera)
Zera uses the same Anthropic Claude models.
Bring your own key (BYOK)
You are able to provide your own API keys. In that case, prompts are sent directly to the provider you configured. Yotzer does not store or log the content of BYOK API calls beyond what is needed for the generation request.
AI-generated content
All content produced by Yotzer tools, including text, images, slides, and video, is AI-generated or AI-assisted.
AI-generated content sometimes contains inaccuracies, hallucinations, or biased outputs. Review every output before you publish or distribute it.
Yotzer does not guarantee the accuracy, originality, or legal compliance of AI-generated content. You are responsible for reviewing and verifying outputs.
How AI content is marked
Yotzer marks AI-generated images and video with a three-layer system:
- Layer 1: a C2PA metadata manifest embedded in image and video outputs, a machine-readable provenance record. Applied on all tiers.
- Layer 2: an imperceptible watermark embedded at generation time. Applied on all tiers. It is a recovery path for when platforms strip metadata on upload.
- Layer 3: a visible "Made with Yotzer" watermark on image and video outputs. Free tier only. Paid tiers are able to remove the visible watermark.
Layers 1 and 2 are the legal compliance floor and cannot be removed on any tier. Only the visible watermark (Layer 3) is tier-gated.
Current status: C2PA tooling is installed, pipeline wiring is in progress. This page will be updated when the marking system is fully operational.
Data handling for AI processing
Prompts sent to AI providers are processed under each provider's data handling policies. Yotzer uses API access, not consumer products, which usually means prompts are not used to train models. See each provider's API terms for details.
For full details on how Yotzer handles your data, see our Privacy Policy.
Your responsibilities
When you use AI-generated content from Yotzer, you agree to the following:
- Do not present AI-generated content as human-created where that distinction is legally required.
- Comply with the laws on AI-generated content in your jurisdiction.
- Obtain any rights or permissions needed for input content, including images, text, and brand materials, that you upload to Yotzer.
EU AI Act compliance
This page is published in compliance with Article 50 of Regulation (EU) 2024/1689, the EU AI Act. It sets out Yotzer's transparency obligations as a deployer of AI systems that use general-purpose AI models.
For AI transparency inquiries, write to privacy@yotzer.app.
Yotzer's full terms are in the Terms of Service.
Updates
This page will be updated as Yotzer's AI capabilities, marking system, or applicable regulations change.