Explore. Create. Go viral.

GOAL
Maximize visual + motion + speech (audio) similarity to the reference video, prioritizing:
1) Subject consistency (identity / wardrobe / props) for the creator in the bottom half.
2) Environment consistency (location, set dressing) for the creator's room.
3) Split-screen layout consistency (Top half: UI/AI Video, Bottom half: Creator).
4) Visual fidelity of the AI-generated casino sequence (lighting, morphing transitions, specific objects).
5) UI accuracy for the screen recording segments.
6) Speech signature (script/semantics, cadence, energetic tone).

WORKFLOW
A) MISE EN PLACE
- Invariants: Split-screen vertical format. Bottom half: Caucasian male, 30s, beard, beige t-shirt, blue baseball cap with yellow logo, sitting in a bright room with a window behind him. Top half: Alternates between high-end cinematic casino AI video and dark-mode web UI screen recordings.
- Variables: The specific content shown in the top half (casino sequence vs. UI), the creator's hand gestures (pointing up, gesturing to explain).

B) SHOTLIST & D) PROMPT SYNTHESIS

MASTER PROMPT:
GLOBAL LOCK: Vertical 9:16 video format featuring a continuous split-screen layout. The bottom half consistently shows a Caucasian man in his 30s with a beard, wearing a beige t-shirt and a blue baseball cap with a yellow logo. He is sitting indoors in a brightly lit room with a window visible in the background. He speaks directly to the camera with an energetic, educational tone, frequently using his hands to gesture upwards towards the top half of the screen. The top half of the screen displays varying content, alternating between a highly polished, cinematic AI-generated video sequence and screen recordings of a dark-mode web interface. The lighting on the man is natural and soft, while the AI sequence features warm, golden, high-contrast cinematic lighting.

[00:00–00:05] Top half: A cinematic, warm-lit shot of a man in a white tuxedo at a casino poker table, pushing chips forward. The camera pushes in rapidly, and the scene seamlessly morphs into an extreme close-up of a hand holding a yellow and black casino chip. Text overlays "Start Frame" and "End Frame" appear briefly, along with a graphic of a video editing timeline. Bottom half: The man points both index fingers upwards at the top screen, looking impressed. Text overlay: "This is AI? 🤯".
[00:05–00:15] Top half: The continuous morphing sequence continues. The yellow casino chip flips into the air, spinning, and seamlessly transforms into a spinning wooden roulette wheel. A white ball drops onto the wheel, bounces, and falls down a dark, golden-lit chute. As it falls, it morphs into two golden dice that land on a green craps table. The dice have the words "Artlist" and "AI Toolkit" engraved on them. In the background, out-of-focus people in formal wear are cheering. Bottom half: The man watches the sequence, resting his hand on his chin, then points downwards, then makes an "okay" hand gesture, reacting to the visuals above.
[00:15–00:35] Top half: The view switches to a screen recording of a dark-mode web interface for "Artlist AI Toolkit". The cursor clicks on "Generate image". The screen scrolls through various AI image models, specifically highlighting "Artlist Original 1.0" and "Nano Banana Pro". Thumbnails of the generated casino images (tuxedo man, chip, roulette) are shown. Bottom half: The man speaks directly to the camera, explaining the process, using his hands to emphasize points.
[00:35–00:44] Top half: The screen recording switches to a video generation interface, highlighting a model called "Kling 2.5 Turbo Pro". The UI demonstrates a workflow using a "Start Frame" and an "End Frame" input, showing the previously generated casino images being placed into these slots to create the transition. Bottom half: The man continues his explanation, gesturing with his hands to indicate the connection between the two frames.
[00:44–00:51] Top half: A static graphic appears showing a dark-themed PDF document titled "GenHQ Workflow Overview" containing thumbnails of the casino images and blocks of text (prompts). A large, bold text overlay appears: "Comment 'AI'". Bottom half: The man points emphatically at the "Comment 'AI'" text, urging the viewer to take action.

NEGATIVE PROMPT:
Visuals: Loss of split-screen format, creator's appearance changing (different hat, different shirt, losing beard), inconsistent lighting on the creator, jarring cuts in the top-half AI sequence, loss of momentum during the morphing transitions, misspelled text on the golden dice (must say "Artlist" and "AI Toolkit"), blurry or illegible UI elements in the screen recording, unnatural hand anatomy during gestures.
Audio: Robotic voice, low energy, muffled audio, lack of sync between creator's lips and speech.

SPEECH PACK:
[00:00-00:15] (No speech, upbeat electronic background music playing)
[00:15-00:35] SPEAKER A (Creator): "Here's how you can make this using Artlist's new AI Toolkit. On Artlist, go to the top of the page and select 'Generate image'. From here, you can go and select all of the best image models, including Artlist's new image model called 1.0. Then, when you pair that with Google Nano Banana Pro, you can insert yourself into the scenes."
[00:35-00:44] SPEAKER A (Creator): "And we'll give you this PDF document at the end of the video with all of our prompts and images. Then you can use the video section of the toolkit, where you can use the best video models, and use Kling 2.5 first and last frame to create these epic transition shots."
[00:44-00:51] SPEAKER A (Creator): "If you want access to all of our prompts and images, type 'AI' in the comments and I'll send you the links. Artlist has a lot of new features coming soon."