#304 Skyway Character Video Workflow

Prompt

Get the Kling video prompt farther below. Use it as you like. I combined a photo I took in the Minneapolis skyway, (Macy’s building) which is the background for the scene, with an image I created using Higgsfield SOUL. I used Seedream for the edit. Then I created a detailed Kling 3 video prompt for the sequence you see here. The prompt is farther below. If you’re not using AI to vibe code video prompts, and targeting them towards specific platforms, give it a try. AI can write an incredible prompt, and it can also tailor it towards your target platform. Each platform has different requirements, and AI has detailed knowledge of what to do. ChatGPT, Google Gemini, Grok, Claude …. Whatever your favorite is will work. I used Seedream to take the image from Higgsfield SOUL (see my previous post for image and image prompt for her) and combine them into one image, and add Black Doc Marten boots. I gave ChatGPT my starting frame for the video, and had it create a detailed Kling prompt for the video for the storyline I had in my mind. 🔥 AI Workflow: AI image creation → Higgsfield SOUL AI image edit → Seedream Image touch up → Photoshop Video prompt creation → ChatGPT Video creation → Kling 3.0 🔥 Kling video prompt: Use the uploaded image as the exact starting frame and exact character identity. Maintain full consistency in face, hair, outfit, proportions, lighting, and environment. Do not alter wardrobe, background architecture, or color tones. Scene: Modern industrial interior with high ceiling and large rectangular overhead light. Clean, minimal, slightly cool color grading. Quiet open-space atmosphere. Camera: Static camera, eye-level. Medium-to-tight framing that allows subject to walk into a close-up. No camera movement. Shallow depth of field, subject always in sharp focus. Action: The woman begins standing at a distance, centered in frame. She walks forward slowly and naturally toward the camera with steady, grounded steps. Her boots produce soft, realistic footstep sounds. As she approaches, her gaze locks onto the lens. Her expression is calm, slightly curious, subtly challenging. She stops about 2 feet from the camera. Final framing shows her full face, neck, and top of shirt clearly. At the end: She leans slightly closer and peers directly into the lens. Her head tilts subtly side-to-side in small, controlled movements, as if trying to see the viewer through the camera. Her expression shifts to a faint, thoughtful curiosity. Performance Details: Natural body movement, no jitter or unnatural motion. Eye tracking remains locked to the lens once she begins walking. Micro-expressions are subtle and realistic. Audio: Minimal sound design. ▪️ Light ambient air tone (open interior space) ▪️ Soft, grounded boot footsteps Timing: 6–8 seconds total duration. Style: Photorealistic, cinematic, natural motion. No stylization, no distortion, no added effects.