NextGen Studio makes YouTube videos and thumbnails for creators. Every request passes through several independent checks before anything is generated, and every image that comes back is checked again before it is saved. This page describes those checks in plain language. The rules themselves are in the Acceptable Use Policy.
1. What we refuse
- Sexual content of any kind, including suggestive framing of any person.
- Anything that sexualises or endangers minors. Reported to NCMEC where U.S. law requires.
- Graphic gore, real-world violence for its own sake, and hateful symbols or slurs.
- Dangerous instructions: weapons, explosives, drug manufacturing, self-harm, fraud.
- The likeness of a real, identifiable person in a thumbnail, and any real person placed in a sexual, humiliating or deceptive scenario anywhere.
- Characters owned by a rights holder, and voice cloning of real people.
Educational, historical, journalistic and fictional treatment of hard subjects is allowed. A history channel can cover a war; it cannot ask for gore.
2. Where the checks run
- Every typed field. Titles, ideas, topics, scripts, scene prompts, animation prompts, character descriptions and style descriptions are screened the moment they are submitted, before any credit is charged.
- The request as a whole. Thumbnail ideas, video topics and editor edits are also read by a classifier that judges what is being asked for, not just which words were used, so rephrasing or another language does not get around it.
- Every prompt sent to a model. The final prompt, whether a person wrote it or our planner did, is screened once more at the moment it leaves for the image or video provider. Trademarked characters are rewritten to generic stand-ins.
- Every image that comes back. Generated thumbnails and scene images are classified by a vision model before they are saved. Generated video clips are sampled and classified the same way. An unsafe result is discarded, not shown.
- Uploads. Reference photos and images are classified before they are stored. Uploaded voiceovers are transcribed and the transcript is screened before any scene is generated from it.
- The providers' own filters sit underneath all of this. We generate through Google's Gemini, Veo and image models and ElevenLabs, each of which applies its own safety policy independently of ours.
3. What happens on a refusal
Nothing is generated and nothing is charged. A refusal for prohibited content counts a strike against the account; three strikes suspend it and alert our team. Refusals for a real person's likeness or an owned character do not cost a strike, since the fix is a reword. Every Thumbnail Studio account agrees to the content rules before its first generation.
4. Reporting
To report content generated on NextGen Studio that violates these rules, email businessvisionmediagroup@gmail.com with a link or description. Intimate or sexual imagery of a real person made without their consent is removed within 48 hours of a valid report. Copyright complaints follow the DMCA process.
5. What this page is not
No filter is perfect, and we do not claim ours is. It is layered so that a miss in one place is caught in another, and it is tightened whenever we find a gap. Operated by Shaun Jonck, trading as NextGen Studio, South Africa.