The Continuity Bible: Same People, Clothes and Places in a Long AI Video
6 min read
By Nureta TeamUpdated November 18, 2026
Yes. Before a Nureta video with more than one scene renders, Nureta writes a continuity bible for it: one fixed plan of the people, what each person wears, and the places. Every shot is planned from it, so the people in the last scene have the same faces, bodies and clothes as in the first, and each place keeps the same background and light. Nureta makes multi-scene videos of up to about 3 minutes with sound. This guide explains what the continuity bible holds, what Nureta will not invent, and 8 steps to write a long story, for example 3 minutes in 8 scenes, that stays consistent.
Nureta is the No.1 unfiltered NSFW AI video platform, trusted by 100,000+ users. Text to video and image to video, up to 3 minutes.
Ready to try? Paste your story and watch it become a video.
Make your first videoThe continuity bible: the facts
Here is what the continuity bible keeps the same, in one table, checked against the app on 2026-11-14.
| Fact (checked 2026-11-14) | Nureta |
|---|---|
| Which videos | Every multi-scene video, longer than 15 seconds |
| What it holds | The people, what each person wears, and the places |
| People | Each person's name, body and face stay fixed for the whole video |
| Where looks come from | Your face photos and your words. Where they disagree, your words win |
| Never invented | An age, a height, a cup size, tattoos, moles or freckles |
| Clothes | A person keeps the clothes of the scene before, unless a scene shows them dressing or undressing |
| Places | One background per place, with the same light, colour temperature and furniture in every scene there |
| Names | The same name for each person in every scene |
| Face photos | Up to 3 on Create, of consenting adults. On the storyboard, each person can have one |
| On the storyboard | The Shared setting panel: People (up to 8), outfits, Locations, and Who, Wearing and Where for each scene |
| Joins in one place | Optional Auto chaining, off by default |
| One wrong section | Fix a part remakes only that section |
What a continuity bible is
On a film set, someone tracks who wears what and where everything stands, so a cut never changes a shirt or moves a lamp. Their notes are often called the continuity bible. A long AI video needs this even more. A video model makes a few seconds at a time, and each clip is a separate generation that remembers nothing of the clip before. Without a shared plan, every gap in the description is filled in again, and a jacket, a tattoo or a room can change between scenes.
Nureta plans the whole video before any shot renders. It reads your story and your photos and writes one plan: who is in each scene, what each person wears when the scene starts, and where it happens. The people are locked: their names, bodies and faces are the same in every shot. Clothes change only when a scene shows it, such as a person taking off a coat, or after a jump in time. Each place gets one background picture with no people in it, a fixed light and colour temperature, and a fixed layout of the furniture, so the bedroom in scene 2 is the same bedroom in scene 7. Every shot is then written and rendered from this plan.
What Nureta will not invent
Appearance comes only from your photos and your words. When the two disagree, for example a photo shows dark hair and you write blonde, your words win. Nureta does not add an age, a height, a cup size, tattoos, moles or freckles that neither your photos nor your words give.
This keeps people stable, but it also means that a detail you leave out is not there. If a tattoo matters, write it and where it is: "a small rose tattoo on her left shoulder". If something must not be there, say so: "no tattoos" is kept as smooth, unmarked skin, and "no jewellery" as a bare neck, ears and wrists. The content policy applies to every scene: adults 18 and older, fictional or consenting, and no real or famous people without their consent.
8 steps for a long video that stays consistent
1. Describe each person once, at the top. Give each person a short block: face, hair, body, skin, tattoos and where they are, clothes, and how they speak. Write it once, before the scenes, and don't describe the same person in other words later: "auburn hair" in one scene and "red hair" in the next can read as two people. For more, read the same face in every scene.
2. Describe each place once. Give each place a name and one line with its light: "Bar: a dim hotel bar, warm amber light, a long wooden counter." Nureta gives each place one background and keeps its light the same in every scene there. On the storyboard, you can also add your own photo of a place.
3. Use the same names every time. Call each person and each place by the same name in every scene: "Anna" in every scene, not "the blonde" in one and "Anna" in the next. Nureta uses the cast names in every part of its plan, so a different word can be read as a different person.
4. Add a face photo for each main person. On Create, add up to 3 face photos of consenting adults, and the same faces are used in every shot. For more people, edit the storyboard: its People list holds up to 8 people, each with a photo of their face or a description of how they look. You can also cast a saved character from your Library. One limit: a shot with your own first or last frame does not use face photos, so choose face photos or frames for a video, not both.
5. Write numbered scenes with camera and dialogue. Start each scene with its number and place, then the camera and the action: "Scene 4 (Room): third-person camera, both in frame. They come in through the door, kissing." For a camera that shows both people, write it in positive words like these, not "not POV". Put each spoken line in curly braces after the speaker's name: Anna whispers {Come closer}. In English, a line in quotes right after a word such as says or whispers also counts. Each line is said once, in the order you write it. To give a person a voice or an accent, open Design their voice on their person card on Create, and every scene asks for that same voice. Read design each character's voice.
6. Keep the scenes that share a place together. Scenes in the same place share one background and one light, and a move to another place is a cut. A story that stays in one place for several scenes in a row, for example scenes 1 to 3 in the bar and 4 to 8 in the room, joins more smoothly than one that goes back and forth. Write clothing changes as actions in the scene where they happen: "Marco takes off his shirt". Otherwise each person keeps what they wore in the scene before.
7. Turn on Auto chaining for smoother joins. For videos longer than 15 seconds, the Auto chaining switch is on the confirm step before you pay, and it is off by default. With it on, each shot in the same place starts from the last frame of the shot before, so the joins are smoother. A change of place is still a cut. It costs nothing extra, but the video takes longer to render. Read Auto chaining.
8. Fix one wrong section with Fix a part. If one section comes out wrong, for example a shirt that changes colour, you don't have to make the whole video again. Fix a part remakes only the clips you pick, and every paid video comes with free fixes. Read free fixes for long videos.
See and change the bible on the storyboard
If you edit the storyboard before you render, the panel Shared setting — the same in every scene shows the continuity bible. Under People you can add a person, up to 8, with a photo of their face or a description of how they look. Removing someone takes them out of every scene. Each person's clothes are listed as outfits. Under Locations you can describe each place and add your own photo of it.
Each scene then has Who, Wearing and Where. The action text decides who is in a scene: name someone in it to add them, and take the name out to remove them. Wearing picks each person's outfit for that scene, and Where picks the place, or Same place as the scene before. Change these before you render, and the whole video follows the new plan. For long videos in general, read the longest AI adult video.
An example: 3 minutes in 8 scenes
Here is a story written with these steps, for a 3-minute video. It is in English, but you can write yours in any language.
People. Anna: woman, late 20s, shoulder-length blonde hair, green eyes, slim build, fair skin, a small rose tattoo on her left shoulder, a black wrap dress, silver earrings, a soft English accent. Marco: man, early 30s, short dark hair, trimmed beard, tall and athletic, olive skin, no tattoos, a white shirt and dark trousers.
Places. Bar: a dim hotel bar, warm amber light, a long wooden counter. Room: a hotel room, white sheets, one bedside lamp, city lights in the window.
Scenes. 1 (Bar): wide shot. Anna sits at the counter. Marco takes the stool next to her and says {Is this seat free?} 2 (Bar): close-up, both in frame. Anna smiles and says in a soft English accent {It is now.} 3 (Bar): medium shot. They talk and laugh, and Anna touches his arm. 4 (Room): third-person camera, both in frame. They come in through the door, kissing. 5 (Room): Marco takes off his shirt and drops it on the chair. 6 (Room): Anna unties her wrap dress and lets it fall. 7 (Room): third-person camera, both in frame. They lie down on the white sheets. 8 (Room): wide shot, the city lights behind them. Anna whispers {Stay.}
Two face photos, one of Anna and one of Marco, would set their faces. Anna's tattoo and earrings are in the plan because the story names them, and the clothes stay off once the story takes them off. With Auto chaining on, the joins in the bar and in the room are smoother, with one cut where the story moves from the bar to the room.