Saudi memory, through a cinematic AI eye. I build the city as I see it in a dream.
What if Gotham were Riyadh?
KAFD towers at sunset, the city on edge.
Joker walks into the bank — shemagh, paint, silence.
Ardha sword dance turns into an UZI.
Batmobile cuts through geometric concrete walls.
A JORER card lands on a Saudi figure.
Final stand: Batman, Kingdom Tower, the city watching.
More journeys coming soon.
Seven portraits, seven mental disorders, one message about awareness.
A Saudi dystopia where rhythm goes extinct.
In a psychiatric ward, a man writes stories no one reads.
An anti-hero who walks out of the screen into the street.
From the 70s artist to the TikTok creator. The evolution of the Saudi creative.
A series of moments without narrative. Visual exercises without endings.
The studio's technical signature. Each character is trained with a custom LoRA model and generated with visual consistency across five angles — preserving identity across films, editorial, and campaigns.
Full visual direction for fashion, advertising, real estate, and heritage — all in the same cinematic language.
"I make Riyadh as I see it in a dream."
Turki Ghazi is a Saudi visual artist and storyteller based in Riyadh, working at the intersection of cinema and artificial intelligence. He builds visual worlds that channel Saudi culture and local memory using the latest generation of generative AI tools.
Since 2025, Turki has produced 9 short films, trained 8 character models at 5 angles each, and directed more than 290 editorial images for brands, real estate developers, and cultural content.
I start from the cultural reference. The film isn't born from a prompt — it's born from a deep reading of Riyadh's scene, popular memory, and symbol.
Written script, sketched storyboard, moodboard for color and lighting. Every shot is defined before I open any tool.
If the project needs a consistent character, I train a custom LoRA model on 40–60 images. The result: visual consistency across multiple angles.
I use Sora, Runway Gen-3, Midjourney, and ComfyUI workflows. Each shot is generated 20–40 times before selecting the keeper.
Edit in Premiere, color grading in DaVinci, sound design crafted scene by scene. A film is not delivered until the sound is sealed.
Currently accepting Q3 2026 projects. Short films, advertising campaigns, music videos, and real estate visualization.