---
title: "AI Movie Generator: Make Cinematic Movie Scenes | Made"
description: "AI movie generator and AI movie maker: cast characters, write the scene and its lines, pick the lens, light and camera move, and get cinematic takes with sound."
url: https://www.made.app/ai-movie-generator
---

[Tools](https://www.made.app/tools) / AI Movie Generator

# AI movie generator: make cinematic movie scenes

Made’s AI movie generator is Cinema Studio, an AI movie maker for cinematic scenes. Cast your characters, write what happens and what they say, pick the lens, the light and the camera move, and get a take of 2 to 30 seconds with sound.

[Open Cinema Studio](https://www.made.app/tools/cinema-studio) See real scenes

Runs in Cinema Studio. You sign in to run it, see the price before you start, and failed runs are refunded.

14 s · 2 shots · 5 lines

What went in

- ![Ren, a man with a bleached buzz cut in a black denim vest](https://www.made.app/images/tools/cinema-studio/ai-movie-generator/cast/ren.webp)
  Ren
- ![Dami, a bearded man with short twists in a mustard cardigan](https://www.made.app/images/tools/cinema-studio/ai-movie-generator/cast/dami.webp)
  Dami
- ![Ivy, a woman with long pink hair and bangs in a white cardigan](https://www.made.app/images/tools/cinema-studio/ai-movie-generator/cast/ivy.webp)
  Ivy
- ![Noor, a woman with a high black ponytail in a white apron](https://www.made.app/images/tools/cinema-studio/ai-movie-generator/cast/noor.webp)
  Noor
- ![Bea, a woman with curly copper hair in a green striped shirt](https://www.made.app/images/tools/cinema-studio/ai-movie-generator/cast/bea.webp)
  Bea
- ![A bright loft with a cream sofa, a glass coffee table and an open kitchen behind it](https://www.made.app/images/tools/cinema-studio/ai-movie-generator/cast/loft.webp)
  The loft

**Cookies for dinner.** Five characters and a loft went in as photos. One take came out: two shots and five spoken lines, with the sound made in the same take. Tap the speaker to hear it. See the script

## Movie scenes, one take at a time

Cinema Studio makes cinematic shots and scenes, not a whole feature film in one go. Each Generate makes one take of 2 to 30 seconds. A take can hold more than one shot, and it grows into a longer scene with Before and Next.

**per take**

2–30 s

per take

**camera moves**

33

camera moves

**shapes, from 9:16 to 21:9**

6

shapes, from 9:16 to 21:9

**photos of your cast, places and props**

10

photos of your cast, places and props

## Make an AI movie scene in 4 steps

1. **Write the scene**
   Say what happens, who is in it and what they say. Add up to 10 photos of the people, places and props, or cast a character you made, and the same person plays every shot.
2. **Direct it**
   Pick a lens, a framing or angle, one of 33 camera moves, the time of day, the weather, the light and a color palette. Each pick is written in as one plain line of direction.
3. **Generate a take**
   Choose the engine, the shape (16:9, 21:9, 9:16 and more) and a length of 2 to 30 seconds. The exact price is on the Generate button, and failed runs are refunded.
4. **Grow it into a scene**
   Press + Next for a clip that opens on the take’s last frame, or + Before for one that leads into its first. Trim, make another version, and Play scene runs the shots in order.

[Write a scene in Cinema Studio](https://www.made.app/tools/cinema-studio)

## Real scenes made in Cinema Studio

Each one is a real take, shown with the words it was typed with and what was picked. Turn the sound on: the sound came out of the same take as the picture.

21:9 · 10 s

### Blue hour on the highway

Words only, with five picks: anamorphic glass, blue hour, practical light, a palette and a dolly in.

Seedance 2.5 · 21:9 · 10 s · words only

- Lens: Anamorphic
- Time of day: Blue hour
- Lighting: Practicals
- Color palette: Ember Roadside
- Camera move: Dolly in

**The words it was made from**

A woman in a long wool coat steps out of a lone roadside diner onto an empty desert highway, the diner's windows glowing behind her. She stops at the edge of the road, pulls her coat tight against the wind, and watches a single pair of headlights approach from far down the highway.

What each pick added

- Anamorphic lenses: a wide cinematic frame, oval bokeh and long horizontal blue streak flares.
- At blue hour: just after sunset, a deep blue sky with the last glow on the horizon and the first lights of the city coming on.
- Practical lighting: the scene is lit only by sources visible in it, such as lamps, signs, candles and screens, with warm pools of light and soft falloff into shadow.
- Color palette: rich and balanced, with shadows in black, highlights in bone, the scene built from maroon, red, apricot and golden yellow. Apply it across the whole frame and keep skin tones believable.
- Dolly in: the camera moves smoothly forward toward the subject on a dolly, so they grow in the frame with a natural change of perspective.

16:9 · 4 s

### Two shots in the rain

Two shots written in the words, a wide shot and a close-up, with nothing picked.

Seedance 2.5 · 16:9 · 4 s · words only

**The words it was made from**

2 shots, with a hard cut from each shot to the next. Shot 1 · 2 s Wide shot: a woman in a long red coat stops in the rain outside a glowing café window at night Shot 2 · 2 s Close-up: her face in the window light as she smiles

The script of the scene at the top

1. 1. Ivy “Are we really eating cookies for dinner?”
2. 2. Ren “We’re being supportive.”
3. 3. Dami “They said twenty minutes. That was an hour ago.”
4. 4. Noor “It’s fine! Everything is fine!”
5. 5. Bea “It is not fine.”

### Cookies for dinner: five characters, five lines

Who sits where, who says which line and how, and the cut at nine seconds are all in the words. Each line came out word for word, in order, from the person it was written for.

Seedance 2.5 · 16:9 · 14 s · 6 photos

**The words it was made from**

2 shots. Total 14 seconds: shot 1 runs 0.0–9.0s, shot 2 runs 9.0–14.0s. Hard cut between them. All real-time, no slow motion. Style: photorealistic live-action film, natural grain, shallow depth of field, pore-level skin. NOT a 3D render, NOT a cartoon. NO ON-SCREEN TEXT — CRITICAL: no captions, subtitles, titles, logos or overlays at any point. THE SCRIPT — CRITICAL: the only words spoken, exactly as written, in this order, each by its assigned speaker. Nothing improvised, added, paraphrased or skipped. 1. @image3: "Are we really eating cookies for dinner?" 2. @image1: "We're being supportive." 3. @image2: "They said twenty minutes. That was an hour ago." 4. @image4: "It's fine! Everything is fine!" 5. @image5: "It is not fine." ONE MOUTH SPEAKS AT A TIME — CRITICAL: each line belongs to exactly one person and only that mouth forms the words. Listeners' mouths stay closed or chewing, never mouthing along. ASSETS (each person matches their reference exactly): @image1 = man, bleached platinum buzz cut, silver ear hoop, black denim vest over a white tee. Voice: dry tenor, deadpan. THIS SCENE: sofa LEFT, eating a cookie. @image2 = man, short twists, neat beard, mustard cardigan over a black tee. Voice: warm baritone, unhurried. THIS SCENE: sofa MIDDLE, holding his phone. @image3 = woman, long bubblegum-pink hair with blunt bangs, white fluffy cardigan, light-blue pleated skirt. Voice: bright mezzo, teasing. THIS SCENE: sofa RIGHT, half-turned to the men. @image4 = woman, long black high ponytail, white chef apron over black long sleeves. Voice: clear alto, loud when stressed. THIS SCENE: at the stove behind the sofa. @image5 = woman, big curly copper-red hair, freckles, green-and-white striped shirt. Voice: low, flat, dry. THIS SCENE: at the kitchen island beside the stove. @image6 = the loft: geography, materials and light only. GEOMETRY MAP: cream sofa in the foreground facing camera, @image1 LEFT, @image2 MIDDLE, @image3 RIGHT, a glass table with a plate of cookies in front. Kitchen island four metres behind: @image4 at the stove screen-left, @image5 at the island screen-right. FIRST FRAME: already mid-scene, all three chewing cookies, the pot behind already boiling over. LENS: shot 1 = 47° (50mm) eye-level three-shot of the sofa, kitchen soft behind. Shot 2 = 63° (35mm) from behind the sofa toward the kitchen, three heads in the foreground. Gentle handheld that settles. ACTION TIMING: 0.0–3.0s (SHOT 1): @image3 turns to the men and says "Are we really eating cookies for dinner?" @image1 and @image2 say nothing in this beat: they keep chewing. 3.0–4.5s: @image1 shrugs and says "We're being supportive." @image2 and @image3 say nothing in this beat: @image3 raises an eyebrow. 4.5–9.0s: @image2 holds up his phone and says "They said twenty minutes. That was an hour ago." @image1 and @image3 say nothing in this beat: they nod. Behind them @image4 stirs the overflowing pot and @image5 chops fast, both silent. 9.0s HARD CUT 9.0–11.5s (SHOT 2): grey smoke puffs from the frying pan; @image4 flaps a towel at it and shouts over her shoulder "It's fine! Everything is fine!" @image5 says nothing in this beat: she stares at the pan. 11.5–14.0s: @image5 says flatly "It is not fine." @image4 says nothing: she keeps flapping. The three on the sofa turn toward the kitchen, silent. AUDIO: diegetic only, bright hard-surfaced room. Line 1 teasing, rising. Line 2 deadpan. Line 3 slow and dry, stress on "hour". Line 4 shouted, too cheerful. Line 5 low and flat. Plus bubbling, a knife on a board, sizzling. No music, no score, no humming, no voices beyond the scripted lines. LOCKS: lines in order 1 to 5, each only from its speaker. Same seating and wardrobe across the cut.

## One scene, three ways

The same run through New York on MiniMax H3 Max: first with nothing picked, then with the Fisheye lens, then with the Fisheye lens and the Snorricam move. Each pick adds one line of direction to the words.

- Nothing picked
  Every setting on Auto: only the words.
  **The words, and what was added**
  woman frantically runs through new york. she bobs and weaves between people. turns around.
- Fisheye lens
  - Lens: Fisheye
  **The words, and what was added**
  woman frantically runs through new york. she bobs and weaves between people. turns around.
  What each pick added
  - Shot on a fisheye lens: an extreme, curved ultra-wide view with strong barrel distortion that bends straight lines into arcs around the center of the frame.
- Fisheye lens and Snorricam
  - Lens: Fisheye
  - Camera move: Snorricam
  **The words, and what was added**
  woman frantically runs throu gh new york. she bobs and weaves between people. turns around.
  What each pick added
  - Shot on a fisheye lens: an extreme, curved ultra-wide view with strong barrel distortion that bends straight lines into arcs around the center of the frame.
  - Snorricam: the camera is rigged to the subject's body, so their face stays locked in the same spot in the frame while the world sways and rushes past behind them.

## 33 camera moves

Pick one move per take and it is written in as its own line of direction. Here are 12 of them, each a real clip from Cinema Studio: a photo Made generated, filmed by WAN 3.0 Prime with the move picked.

- Dolly in
  Move forward toward the subject
- Crane up
  Rise on a crane arm
- Drone orbit
  A wide circle from the air
- Dolly zoom
  Push in while zooming out
- Whip pan
  A fast pan with motion blur
- Snorricam
  Face locked, world in motion
- Side tracking
  Travel alongside in profile
- Rack focus
  Focus shifts from near to far
- Aerial pullback
  Lift and reveal from above
- Bullet time
  Frozen moment, camera sweeps
- Tilt down
  Pivot down onto the subject
- Helicopter shot
  A high, floating aerial move

## Everything you can direct

Every setting starts on Auto. Pick what matters to the shot and leave the rest: each pick becomes one plain line of direction after your words.

### Lens

From a 14mm ultra wide to a 135mm telephoto, a fisheye, or anamorphic and vintage glass.

- 14mm
- 24mm
- 35mm
- 50mm
- 85mm
- 135mm
- Fisheye
- Anamorphic
- Vintage Anamorphic
- Warm Vintage
- Halation Vintage

### Framing and angle

10 framings, from a wide shot to an extreme close-up, or one of 7 camera angles.

- Wide
- Medium
- Close-up
- Extreme close-up
- Two-shot
- Over the shoulder
- POV
- Selfie
- Insert
- Through the foreground
- Low angle
- Worm's-eye
- Ground level
- High angle
- Top-down
- Aerial
- Dutch angle

### 33 camera moves

One move per take, from a dolly in to a drone orbit, a whip pan or bullet time.

- Snorricam
- Robot arm
- Tilt up
- Rack focus
- Tilt down
- POV
- Pan left
- Crane up
- Pan right
- Crane down
- Side tracking
- Pedestal up
- Pedestal down
- Handheld
- Tracking
- Drone orbit
- Dolly zoom
- Aerial pullback
- Static shot
- Bullet time
- Whip pan
- Slow zoom in
- Arc left
- Slow zoom out
- Arc right
- Truck right
- Dolly in
- Truck left
- Dolly out
- Slider right
- Crash zoom
- Slider left
- Helicopter shot

### Time, weather and speed

Dawn, golden hour, midday, blue hour and night; clear, overcast, rain, fog and snow; slow motion to a hyperlapse.

- Dawn
- Golden hour
- Midday
- Blue hour
- Night
- Clear
- Overcast
- Rain
- Fog
- Snow
- Slow motion
- Speed ramp
- Fast motion
- Time-lapse
- Hyperlapse

### Light and color

6 lighting setups or up to two lights of your own, and 50 color palettes, or one made from a photo.

- Silhouette
- Practicals
- Window
- Overhead fall
- Contre jour
- Soft cross

## Engines, lengths and shapes

Pick the engine on the model chip under the prompt. The settings change with it: each one shows only what that engine has.

| Engine | Takes | Length | Shapes | Sizes | Sound |
| --- | --- | --- | --- | --- | --- |
| WAN 3.0 Prime | Words only, up to 10 photos of the people, places and props or a first and last frame | 2 to 30 s | 16:9, 4:3, 1:1, 3:4, 9:16 | 480p, 720p, 1080p | Yes, made with the picture |
| Seedance 2.5 | Words only or up to 9 photos of the people, places and props | 4 to 30 s | 1:1, 3:4, 9:16, 4:3, 16:9, 21:9 | 480p, 720p, 1080p | Yes, made with the picture |
| Kling 3 Pro | Words only or a first and last frame | 3 to 15 s | 16:9, 9:16, 1:1 | One size | On or off, a switch |
| MiniMax H3 Max | Words only or a first frame, a last frame, or only a last frame | 5 to 15 s | 1:1, 3:4, 9:16, 4:3, 16:9, 21:9 | 480p, 768p, 1080p | Yes, made with the picture |

Each Generate makes one take, and the price is on the button before you run it. Your projects and takes are kept with your Made account.

## Questions about the AI movie generator

**What is an AI movie generator?**

It turns a written scene into film: the people, the place, the action and the camera. On Made it is Cinema Studio. You write the scene, cast your characters and pick the lens, light and camera move, and it films a take with sound.

**Can it make a full movie?**

Not in one go. Each Generate makes one take of 2 to 30 seconds, so a movie is built a scene at a time. Grow a take with + Next (it opens on the last frame) and + Before (it leads into the first), and Play scene runs the shots in order. Download each take to cut them together in any video editor.

**Is there a free AI movie generator on Made?**

No. Made has no free generations: every take is paid, and the exact price is on the Generate button before you start. You can open Cinema Studio, write a scene and set it up without paying, and failed runs are refunded.

**How long can a scene be?**

A take runs 2 to 30 seconds, depending on the engine: up to 30 on WAN 3.0 Prime and Seedance 2.5, up to 15 on Kling 3 Pro and MiniMax H3 Max. A scene is as many takes as you add with Before and Next.

**Can I keep the same characters in every shot?**

Yes. Cast a character you made in the AI Influencer studio and their photo goes with every take, so the same person plays each shot. Seedance 2.5 takes up to 9 photos of people, places and props, like the five friends and the loft above. Use only yourself, people who agreed, or characters you made: Seedance asks you to confirm you are 18 or older and have their permission.

**Does it make sound and dialogue?**

Yes. WAN 3.0 Prime, Seedance 2.5 and MiniMax H3 Max make the sound with the picture, and Kling 3 Pro has a switch for it. Write the lines in your scene and Seedance 2.5 speaks them, as in the scene at the top of this page.

**Which aspect ratios can I make?**

1:1, 3:4, 9:16, 4:3, 16:9 and 21:9. The wide 21:9 frame is on Seedance 2.5 and MiniMax H3 Max; vertical 9:16 is on every engine, for Reels, TikTok and Shorts. Sizes go from 480p up to 1080p, depending on the engine.

**What camera controls are there?**

33 camera moves, from a dolly in to a drone orbit, lenses from a 14mm ultra wide to a 135mm telephoto, a fisheye, anamorphic and vintage glass, 10 framings, 7 camera angles, the time of day, the weather, slow motion, 6 lighting setups and 50 color palettes.

**How much does it cost?**

Video is priced by the second, so a longer take costs more. The exact price is on the Generate button before you run anything, and failed runs are refunded.

**Is it different from a text to video generator?**

It runs the same kind of video models, with a director’s controls on top: a cast that stays the same, a lens, an angle, a camera move, light and color, and takes that grow into scenes. For one quick clip from a prompt, the AI video generator is simpler.

[Open Cinema Studio](https://www.made.app/tools/cinema-studio) [Browse every tool](https://www.made.app/tools)

## More ways to make video

- [**AI video generator**](https://www.made.app/ai-video-generator)
  One clip from a prompt, on every top video model.
- [**AI storyboard generator**](https://www.made.app/storyboard-generator)
  Plan the shots as boards before you film them.
- [**Movie poster maker**](https://www.made.app/movie-poster-maker)
  A poster for your film, with you on it.
- [**Text to video**](https://www.made.app/text-to-video-generator)
  A five-second clip straight from a sentence.

---

Markdown version of https://www.made.app/ai-movie-generator. Every public Made page has one at its URL plus .md (https://www.made.app/ai-movie-generator.md). Site summary for AI assistants: https://www.made.app/llms.txt
