Your photos become cinematic characters. Your words become dialogue. Four AI tools, one finished film — starring you.
Snap a photo of each volunteer. These become the character reference images that every AI tool uses to keep the face consistent.
Upload the photo to ChatGPT and ask it to render the person as a cinematic character in our chosen style — film noir, Wes Anderson, classic Hollywood, whatever fits the script. ChatGPT produces better character portraits than Kling's own image generator — that's why we use it for this step.
Upload each character image to Kling's Element Library and enable Bind Subject. This locks the face in 3D so it stays identical across every scene — no drifting, no morphing.
Use Image-to-Video mode with the Omni model. Write the scene description and dialogue directly in the prompt. Kling generates 3–7 second video clips with native lip sync — the character speaks the lines. Audio, voice, and video all rendered in a single pass.
Import all clips into CapCut. Arrange in order, add transitions, music, titles, credits. Export the finished film.
Generates character images from your photos. Upload a real photo → get a cinematic render in any visual style you choose.
Turns still images into video clips with native dialogue and lip sync. Write what they say — Kling makes them say it. 3–7 seconds per clip. Use the Omni model — it's the one with built-in audio and lip sync.
Free video editor. Drag clips in order, add transitions, music, and titles. Export the finished film in one click.
Wrote both scripts. Can rewrite dialogue, adjust tone, or create entirely new scenes on the fly during class.
We need 2–3 volunteers for each film. Each person gets their photo taken — that photo becomes their character.
⚠️ Privacy note: Photos are uploaded to AI services. Only use photos of people who've given permission.
Inspired by Fawlty Towers. A hotel manager in magnificent denial about a cat. Comedy.
Needs: 2 actors — the Manager and the Hotel Guest
Inspired by Seinfeld. Two people forensically analyzing the meaning of waving. Comedy.
Needs: 2–3 actors — the Philosopher, the Analyst, and the Walker outside
Upload each volunteer's photo to ChatGPT and ask it to render them as a cinematic character.
Kling 3.0 has its own image generator, but ChatGPT produces more detailed, prompt-faithful character portraits — especially for "render this real person in a movie style." We ask ChatGPT to generate 3 separate angles of each character (frontal, side, three-quarter) so Kling's Element Library gets the multi-angle references it needs to lock the face in 3D.
Step 1: Paste with the volunteer's photo attached. Steps 2 & 3: paste in the same chat after the first image generates.
Prompt 1 — Frontal: Using the attached photo as a face and body reference, create a cinematic portrait of this person as a flustered British hotel manager in the 1970s. Frontal view, standing behind a hotel front desk, strained smile, hands clasped. White dress shirt with sleeves slightly rolled, loosened tie, dark waistcoat. Small British hotel lobby — wooden front desk, brass bell, key rack on the wall, faded floral wallpaper. Warm tungsten lighting, slight film grain, Wes Anderson color palette with muted teals and warm yellows. Photorealistic, cinematic aspect ratio.
Prompt 2 — Side profile: Now generate the same character from a side profile view. Same outfit, same setting, same lighting and style. Looking down at papers on the desk, composure tightly held.
Prompt 3 — Three-quarter: Now generate the same character from a three-quarter angle. Same outfit, same setting, same lighting and style. Leaning slightly forward over the desk, one eyebrow raised.
Step 1: Paste with the volunteer's photo attached. Steps 2 & 3: paste in the same chat.
Prompt 1 — Frontal: Using the attached photo as a face and body reference, create a cinematic portrait of this person as an unimpressed hotel guest in the 1970s. Frontal view, standing at a hotel front desk, arms folded, patient but skeptical expression. Smart cardigan and reading glasses. Small British hotel lobby — wooden desk, brass fixtures, faded floral wallpaper. Warm tungsten lighting, slight film grain, Wes Anderson color palette. Photorealistic, cinematic aspect ratio.
Prompt 2 — Side profile: Now generate the same character from a side profile view. Same outfit, same setting, same lighting and style. Looking toward the clerk with one eyebrow slightly raised.
Prompt 3 — Three-quarter: Now generate the same character from a three-quarter angle. Same outfit, same setting, same lighting and style. Leaning on the front desk counter, the look of someone who absolutely did see a cat.
Prompt 1 with photo attached. Prompts 2 & 3 in the same chat.
Prompt 1 — Frontal: Using the attached photo as a face and body reference, create a cinematic portrait of this person sitting in a classic American diner booth. Frontal view, leaning forward with the intensity of someone who just asked an extremely important question. Coffee cup in front of them. 1990s setting — red vinyl seats, Formica table, window showing a city sidewalk. Casual clothes — button-down shirt, maybe a jacket. Warm natural light from the window, slight film grain, Seinfeld-era New York diner aesthetic. Photorealistic, cinematic aspect ratio.
Prompt 2 — Side profile: Now generate the same character from a side profile view. Same outfit, same diner setting, same lighting and style. Gesturing with one hand while making a point, window light on their face.
Prompt 3 — Three-quarter: Now generate the same character from a three-quarter angle. Same outfit, same diner setting, same lighting and style. Chin slightly raised, mid-sentence, animated expression.
Prompt 1 with photo attached. Prompts 2 & 3 in the same chat.
Prompt 1 — Frontal: Using the attached photo as a face and body reference, create a cinematic portrait of this person sitting across a diner booth from someone. Frontal view, listening with genuine concentration, chin in hand. 1990s American diner — red vinyl seats, Formica table, coffee. Casual clothes. Warm window light, slight film grain, Seinfeld-era diner aesthetic. Photorealistic, cinematic aspect ratio.
Prompt 2 — Side profile: Now generate the same character from a side profile view. Same outfit, same diner setting, same lighting and style. Coffee cup raised halfway, thoughtful expression.
Prompt 3 — Three-quarter: Now generate the same character from a three-quarter angle. Same outfit, same diner setting, same lighting and style. One eyebrow raised, about to respond to something surprising.
Prompt 1 with photo attached. Prompts 2 & 3 in the same chat.
Prompt 1 — Frontal: Using the attached photo as a face and body reference, create a cinematic image of this person walking past a diner window on a city sidewalk. Frontal view, smiling broadly and waving enthusiastically with both hands raised — a big, joyful, two-handed wave. 1990s setting, casual clothes, daylight, slight film grain. Photorealistic, cinematic aspect ratio.
Prompt 2 — Through the glass: Now generate the same character from a three-quarter angle, shot from inside the diner through the glass. Same outfit, same daylight, same style. Mid-wave, full joy on the face.
Prompt 3 — Side profile: Now generate the same character from a side profile view. Same outfit, same style. Arms still up mid-wave, walking past the window, caught in the act of an enormous friendly gesture.
Before generating any video, we create a Character Element in Kling's library for each actor. This locks their face, clothing, and voice so they stay identical across every scene.
Go to klingai.com. Choose Kling Video 3.0 Omni — this is the only model with native audio and lip sync. The standard 3.0 is silent.
Click Elements in the sidebar, then Create New Element. Choose Character as the type.
Upload all 3 images ChatGPT generated for this character — frontal, side profile, and three-quarter angle. Use the frontal as the Main Reference and the other two as supplementary. Kling uses these multiple angles to build a 3D map of the face, which keeps it stable across camera moves.
Give the element a clear name — "Manager," "Hotel Guest," "Philosopher." Kling auto-generates a descriptive prompt based on the uploaded images (what the character looks like, what they're wearing, etc.). Review it — you can edit it, but the auto-description is usually good.
Kling asks you to bind a voice to the character. You have three options:
Save the element. Repeat for each character. Once saved, these live in your library and can be reused across every scene — you don't re-upload each time.
Without Element Binding, Kling can "drift" — the character's face changes between clips. With it, the same person looks and sounds like the same person in every scene. This is what makes it a film instead of random clips.
When generating a video clip:
Film A: @Manager and @Hotel Guest
Film B: @Philosopher, @Analyst, and @Walker
Name your elements exactly this way — the prompts reference them.
Kling gives you 66 free credits every day — no credit card needed. Enough for several clips. Credits reset daily but don't roll over.
A hotel manager. A missing cat. A magnificent refusal to acknowledge reality. Inspired by Fawlty Towers.
INT. HOTEL LOBBY — DAY. The Guest approaches the front desk. The Manager stands behind it, straightening papers.
Same location. The Manager hasn't moved. The Guest is patient but firm.
The Manager returns from upstairs, straightening a tie with perfect composure.
A faint meowing sound from upstairs. Both hear it. Only one acknowledges it.
Another meow — louder this time. A long, unmistakable purr.
The Manager reaches under the desk with great dignity and produces a small printed card.
The Manager slides it across the desk. The Guest stares at it. From upstairs: one final, triumphant meow. The Manager doesn't blink. Fade to black.
Use Kling Video 3.0 Omni · Image-to-Video mode · Bind elements · Enable audio · Enable Multi-Shot. Copy buttons grab the full prompt.
Start Frame: Hotel Guest image · Bind @Manager + @Hotel Guest · 3 shots
Guest arrives, reports the cat, Manager denies. Escalates through "We don't allow cats" / "And yet."
🔗 Open KlingStart Frame: Manager image · Bind @Manager + @Hotel Guest · 2 shots
Manager returns with the pillow diagnosis. "A small decorative pillow…" / "The pillow was purring." / "Some of our pillows are very high quality."
Start Frame: Manager image · Bind @Manager + @Hotel Guest · 3 shots
Meowing from upstairs. "What was that?" / "The pipes." / "Victorian plumbing. Very vocal." Manager straightens papers, perfectly calm.
Start Frame: Manager image · Bind @Manager + @Hotel Guest · 3 shots
The finale. "Twenty-three years…That is plumbing." Comment card. Final meow. Manager does not blink.
Each prompt uses Kling's multi-shot syntax: Shot 1 (5s): description. Shot 2 (5s): description. Kling handles the camera cuts automatically. Enable Multi-Shot mode before pasting. CapCut stitches the clips into the final film.
Two people in a diner, forensically deconstructing the meaning of a wave. Inspired by Seinfeld.
INT. CLASSIC AMERICAN DINER — DAY. Two people sit across from each other in a booth. Coffee cups. Window beside them.
Same booth. They're warming up.
B raises a finger to make a point.
A person walks past the diner window outside. They spot A and B, and wave — a big, full, enthusiastic, two-handed wave. Both A and B freeze mid-sip.
They slowly return to their coffee. The mood has shifted — they've been shaken.
They both stare into their coffee cups. Fade to black.
Same setup: Kling Video 3.0 Omni · Image-to-Video · Bind elements · Enable audio + Multi-Shot · Element names: @Philosopher, @Analyst, @Walker.
Start Frame: Philosopher image · Bind @Philosopher + @Analyst · 3 shots
Diner establishing shot. "Do you think people who wave with both hands are happier?" / "That's a really good question." / "That's a person who is all in on the greeting."
🔗 Open KlingStart Frame: Analyst image · Bind @Philosopher + @Analyst · 3 shots
"Those people are hiding something." / "I'm not saying they're criminals. I'm saying I wouldn't lend them my car." Analyst laughs. Philosopher stays serious.
Start Frame: Philosopher image · Bind @Philosopher + @Analyst · 3 shots
"That's not a wave. That's a dismissal." / "Are you the queen?" / "Sometimes at brunch I feel like the queen."
Start Frame: Walker image · Bind @Walker + @Philosopher + @Analyst · 3 shots
Walker's two-handed wave through the window. Cut to interior — both freeze. "Did you see that?" / "Both hands." / "Full extension." / "That was aggressive."
Start Frame: Philosopher image · Bind @Philosopher + @Analyst · 2 shots
Staring into coffee. "I feel like we need to talk about what just happened." / "I don't think I'm ready."
4 clips · ~52 seconds total
5 clips · ~70 seconds total
Each prompt uses Kling's multi-shot syntax: Shot 1 (5s): description. Shot 2 (5s): description. Enable Multi-Shot mode before pasting. Kling handles the camera cuts between shots automatically. CapCut stitches the clips into the final film.
Save every Kling clip to your phone or computer. Name them in order: Scene1.mp4, Scene2.mp4, etc.
Create a new project. Import all scene clips. Drag them onto the timeline in script order.
Use simple cuts between most scenes. Add a fade to black before the final scene for dramatic effect.
First frame: film title in a clean font over black. Last frame: "Made with AI — Class 5" credit.
CapCut has a free music library. A light background track can tie the scenes together. Keep it quiet under dialogue.
Tap Export → 1080p. Your finished film downloads in seconds. Share it, show it, keep it.
If a scene isn't working, ask Claude to rewrite the dialogue shorter or with different phrasing. Then try again. The script is never locked — revision is instant and free.
A short film — with your faces,
real dialogue, and cinematic visuals —
made entirely by AI in one class.
Made in class. Photos of real people. AI-generated dialogue, lip sync, and video. Stitched in CapCut. Here they are.
A hotel manager in magnificent denial about a cat — made entirely with AI. Student photos became cinematic characters. Every line of dialogue was generated with lip sync by Kling 3.0 Omni. Assembled in CapCut.
📸 Phone photo → 🎨 ChatGPT (3 angles) → 🔒 Kling Element Library → 🎬 Kling 3.0 Omni (4 multi-shot clips) → ✂️ CapCut
Two people in a diner, forensically deconstructing the meaning of a wave — then getting ambushed by one. Student photos, AI dialogue, AI lip sync, AI video. Made in class.
📸 Phone photo → 🎨 ChatGPT (3 angles) → 🔒 Kling Element Library → 🎬 Kling 3.0 Omni (5 multi-shot clips) → ✂️ CapCut
Everything we've learned across five classes — combined into one final project.
🎬
You are now AI filmmakers.
No, seriously. You just made a movie.
1 more class to go.
adam@reclamation.ai