Contentrue is built on AI. This note explains where we use it, how we label what it makes, and what you can do to be open with your own audience. It supports the transparency duties that apply to synthetic media, including under the EU AI Act (Article 50) for customers who publish in the EU.
1. Where AI is used
- Planning and writing. Language models from OpenAI and Google (and Anthropic, where used) plan scenes, write captions and, when you ask for it, suggest topics.
- Images. Image models from OpenAI and Google create the visuals. A vision model checks results against your continuity lock and flags mismatches.
- Voice. Narration is synthetic. Turkish voices are generated by open models running on our own servers; other languages use cloud text-to-speech from OpenAI or Google. All voices are generic synthetic voices. We do not offer cloning of real people's voices.
- Video. Reels are assembled from your approved images and narration with conventional video software (not generative video).
AI providers handle your prompts under their API or business terms. We do not use your content to train models, and we do not let providers do so by default. See the Subprocessors page.
2. Human control
AI never publishes on its own. You (or a team member or client you give a link to) review and approve the plan and the final post. Every approval and publish is recorded in the audit log.
3. How synthetic content is labelled
- Synthetic narration. Recipes include a disclosure line that is added to the end of the caption of videos with generated voice (for example "The voice in this video is AI-generated"). It is on by default; you can edit or turn it off, and then you are responsible for any disclosure required where you publish.
- Image metadata. Images from some providers carry content credentials (C2PA metadata). Social networks may read these and add their own label, such as Meta's "AI info". We do not strip this metadata.
- Platform labels. Instagram and X may offer or require their own AI-content labels. Use them where they apply.
4. Your responsibilities
- Check facts. AI can be wrong, make up details or reflect bias. Fact-based recipes include an accuracy check, which reduces but does not remove the risk.
- Do not present synthetic media as real footage of real people or events, and do not create deepfakes of real people (see the Acceptable Use Policy).
- Follow advertising and platform disclosure rules that apply to you, for example labelling AI-generated or manipulated content that could be mistaken for genuine.
5. Limits you should know about
AI output may resemble other people's output and may not be protected by copyright in all countries. Voices are not tuned to individual speakers, may mispronounce foreign words (use the pronunciation dictionary), and cannot convey every emotion. Images may contain small errors, especially in text and hands.
6. Questions or concerns
If you see Contentrue output that you believe is harmful or misleading, tell us at hello@contentrue.com.