Created with intention and systems orchestration.Source materials, prompt engineering, tool knowledge,producing high fidelity results.
One In 10,000
Original narrative developed into a complete AI-assisted filmwith Claude, ComfyUI, and a DaVinci Resolve finishing pass.
Creative Direction / Workflow
Story and concept developed by Jeremy Martin. Claude translated the narrative direction into production prompts and generated the voiceover and music. Visual sequences were assembled through a custom ComfyUI workflow, then edited, mixed, and finished in DaVinci Resolve.
Storyboard Development
Concept Board v01Refined Board v02
Final FilmClaude · ComfyUI · DaVinci Resolve
Girls Painting Motion Study
Original digital painting revived through image-to-videoprompting, preserving the painterly surface, pose language, and composition.
Prompt
Two girls in a soft painterly portrait come alive with subtle, happy, playful motion. Gentle smiles grow, a light shared glance toward the viewer, small natural shoulder and head movement, delicate pose shifts, relaxed crossed-arm attitude preserved, soft hand motion only, no limb warping, no abrupt hand jumps, maintain the original painted texture, brushwork, colors, and composition, cinematic but very subtle animation.
Source Painting
Final Motion OutputFlora AI · Image to Video
Wizard Girl Motion Study
Character image translated into restrained cinematic motionwhile preserving pose, wardrobe, atmosphere, and composition.
Prompt
Wizard girl in a cinematic fantasy portrait comes alive with subtle natural motion. Gentle breathing, small head and shoulder movement, soft fabric motion, calm confident presence, preserve the original character design, wardrobe, lighting, color palette, and composition, no face warping, no abrupt limb motion, no distorted hands, refined magical atmosphere, elegant image-to-video motion.
Source Images
Original PaintingLuma Source
Final Motion OutputLuma AI · Image to Video
Jellyfish Iteration Study
Motion development shown from rejected first passesthrough the third render selected as the final output.
Prompt / Direction
Animate the jellyfish with elegant underwater movement, slow organic drifting, soft translucent body motion, and subtle luminous atmosphere. Preserve the source image composition, color, and delicate texture while refining the motion through multiple generated passes.
Source Image
Final Motion OutputLuma AI · Image to Video
Pass 01 · Rejected
Pass 02 · Rejected
Pass 03 · Accepted
Cherry Blossoms Motion Study
Still image translated into atmospheric botanical motionwith delicate petal movement, soft depth, and restrained cinematic pacing.
Prompt / Direction
Animate cherry blossoms with graceful natural movement, soft breeze-driven petal motion, subtle parallax depth, gentle organic camera energy, and a calm cinematic atmosphere. Preserve the original composition, color, delicate branch structure, and painterly softness while creating a refined image-to-video render.
Source Image
Final Motion OutputLuma AI · Image to Video
Time Unwound Motion Study
ChatGPT-generated character art for the book Time Unwound animatedwith Flora AI, controlled pose, linework, and vintage illustration texture.
Prompt / Direction
Animate the illustrated Harold J. Immerman character art with subtle, intentional motion suitable for a book world treatment. Preserve the vintage ink-and-paper illustration style, typography, suit details, glasses, and full-body pose while adding restrained cinematic movement and a polished image-to-video finish.
Source Image
Final Motion OutputFlora AI · Image to Video
Dijon 1896 Motion Study
ChatGPT-generated historical market square artwork animatedwith Gemini Veo 3.1, preserving atmosphere, period detail, and painterly texture.
Prompt / Direction
Bring the 1896 Dijon market square scene to life with the sweeping motion of a cinematic fly over. Preserve the ChatGPT-generated painterly texture, old-world architecture, signage, crowd density, and warm period atmosphere while adding restrained street movement, ambient depth, and natural historical realism.
Source Image
Motion OutputsGemini Veo 3.1 · Image to Video
Veo 3.1 · Pass 01
Veo 3.1 · Pass 02
Personal Project · Vibe Coded with AI
RoleLens
Executive Decision Engine
Built on evenings and weekends, with AI as co-developer. Not a client project, not a work deliverable. A personal tool I'm actively building toward eventual release on iOS and Android. I brought the product vision, the UX thinking, and the instinct that this thing should exist. The AI handled the code. The result is something that actually works: paste any job posting from LinkedIn, Indeed, or Glassdoor and RoleLens decodes it against your personal risk appetite, life anchors, and career stage, with compensation data sourced from BLS, MIT Living Wage, Glassdoor, and Levels.fyi. Still in active development. This is what vibe coding looks like when you bring real product instincts to the session.
A meditative puzzle game where you restore broken timepieces and uncover the stories held within them. Each watch carries a history, a timeline, and a purpose. Built entirely through AI collaboration with Claude, from game design and narrative structure to every line of React. Speak with AION, an ancient intelligence who has been waiting in the workshop for you to arrive. Enjoy the fully playable game below.
Interactive demo blocked by browser security.
PlayableFull game experience. Click into the frame to begin.
