On this page you will find the prompts, graybox video, and environmental reference (and the AI video tutorial!). Remember to add a character sheet or photo. You can also download this ZIP file.

Generate the video on OpenArt and use this discount code for 15% off: AIVS15

CHATGPT / BLENDER REVIEW ROOM PROMPT

Use this prompt to get ChatGPT to create pre-visualization animations. Make sure you have Blender installed and are using ChatGPT’s desktop Work tab with “On your computer” selected:

I want to create a reviewable previsualization for an AI-generated video. I have Blender installed on this computer.

Story: [Describe the action in a few sentences.]
Target length and format: [For example, 20 seconds, 16:9]
Visual references: [List any storyboard, character, location or prop files.]
Essential geography and physical constraints: [Who or what moves where? What must remain connected, open, visible or separate?]
Key action beats: [List the moments that must happen in order.]

Build a simple Blender previz with a continuous camera path and readable timing. Prioritize staging, motion, scale, physical contact and clear sightlines over realistic rendering. Make the moving elements visually distinct so I can tell what each shape represents. Include an action version with a simple character proxy and an environment-only version using the same camera and timing.

Render both previews and create a local browser review room where I can play and scrub them, step frame by frame, switch versions, attach notes to exact frames, and download the current frame, videos and feedback. Keep editable source files and each revision in a clearly named folder. Open the review room when it is ready.

Before calling the previz finished, check the key events and likely continuity failures against the rendered frames. When I leave feedback, revise the scene and show me the updated preview. Later, help me choose a small set of reference frames and write prompts for my image and video models using their official documentation.

GRAYBOX ANIMATION / PREVIZ (Environment Only Camera Motion)

(click the three dots in the lower right to download it)

SEEDANCE 2.5 ALPINE ESCAPE PROMPT

Generate a cinematic, photorealistic live-action Alpine escape in one continuous 30-second, 16:9 shot.

Reference roles:

@video1 controls camera movement, framing, timing, character blocking, spatial relationships, gondola trajectories, cable failure and impact. Follow its exact event timing and continuous camera path. The whole-second ranges below describe action phases. Use its animation as staging; rebuild the simplified scenery and surfaces as a photographed real environment. Throughout the video, the terrain should appear natural and not replicate the sharp geometry of the previz video.

@image1 supplies the identity and clothing of the single woman. Replace the previz mannequin with her, maintaining her identity and outfit throughout.

@image2 supplies the Alpine environment, avalanche appearance, daylight and photographic treatment. It also supplies the graphite gondola finish, glass, roof and rail details. Apply this appearance consistently to both cabins. 

Her performance is driven by fear, adrenaline and determination to survive. She reacts to each new threat: eyes tracking the approaching cabin, a sharp breath before the leap, a startled gasp as her grip catches and her weight drops. While dangling, her gripping arm trembles; she looks urgently for another handhold, then toward the sound of the straining cable. During the slide, she squints against the spray, tracks the approaching pole, glances back at the pursuing cabin and commits to braking with clenched determination. Her movements are urgent and imperfect, with visible effort. Once inside, she stays braced against the cabin’s jolts as the avalanche pushes it downhill. She watches the terrain passing through the open hatch, eyes searching ahead. She releases a shaky breath, registering that she survived the fall, while remaining tense and ready to react. Keep these reactions natural and integrated into the continuing motion, without pausing the choreography.
Match the environment reference’s lighting and photographic color treatment. Believable skin and fabric, subtle glass reflections, restrained metal wear and soft natural snow surfaces. The avalanche is a towering, connected wall of snow descending from the visible mountain, rising above the cabins and cable supports. Its dense rolling front overtakes the cable infrastructure, with broad billowing powder above a heavy moving snow mass. Translate the rounded avalanche placeholders into this continuous advancing front. Keep airborne particles restrained and surfaces naturally detailed. Preserve the functional hatch openings and connected handrails.

Two matching resort gondolas travel in opposite directions on separate parallel cables. Her original cabin moves toward the avalanche; the escape cabin moves away.

0–5s: Close outside her window, with her back to the camera as she watches the approaching avalanche through the cabin. She turns, looks up, spots the roof hatch and urgently climbs through it. Track upward beside her, revealing the escape cabin, the mountain slope and the towering avalanche approaching along the cable route.

5–8s: She jumps across the gap between the passing roofs. During her jump, the uphill support is swallowed by the avalanche and the cable runs disappear into the advancing snow. Her waist and hip strike the escape cabin’s upper side, just below the roof edge. The impact knocks the breath from her and her body drops down the outside of the cabin. One hand catches the connected roof rail, abruptly arresting her fall. Follow her downward as her weight loads the gripping arm and her legs swing against the glass.

8–13s: She swings beneath her gripping hand and reaches unsuccessfully with her free hand. Her original cabin disappears into the avalanche behind her. Hanging from the roof rail with one hand, she repeatedly tries to pull herself up. Her gripping elbow bends under the effort, but she cannot raise herself over the edge. She draws her knees toward the cabin and presses her boot soles against the window, searching for enough friction to push upward. Her boots slip down the smooth glass, dropping her weight back onto her straining arm and swinging her body outward. Her free hand searches urgently for another hold. Preserve the reference framing that makes its engulfment visible. Follow the camera’s developing reveal of the strained suspension and loaded cable.

13–17s: At the moment shown in the video reference, the taut load-bearing cable snaps. Its broken ends recoil and fall slack as the cabin loses support. Follow her accelerating descent as she releases her grip and falls below the cabin. Hold the suspenseful interval with the landing surface out of view, then reveal the steep powder slope approaching beneath her. Preserve the separation and falling trajectories shown in the reference.

17–19s: She makes contact along the steep downhill powder slope, carrying her momentum down the incline through one somersault rather than stopping on impact. The cabin strikes the slope behind her and rotates onto its side, its roof opening facing downhill. Her roll resolves into a seated, boots-first slide, leaning backward with her boots pointing downhill toward the retreating camera.

19–21s: Immediately after the roll, she uses her hands to steer narrowly past the fixed downhill support pole. The following cabin passes the same pole: its projecting open hatch lid strikes the pole and tears free. The detached lid flies screen right, tumbles back uphill relative to the continuing cabin, then disappears into the pursuing avalanche. Preserve the reference’s contact point and debris trajectory. The woman clears the pole without colliding with it.

21–23s: Continue her seated, boots-first downhill slide. Keep the woman in the foreground, the cabin following uphill behind her and the avalanche behind the cabin. Maintain the following gap until she deliberately brakes. She glances back and aligns herself with the now-unobstructed roof opening.

23–25s: She plants both boots to brake. The cabin catches up as she slows, and she tucks backward through the roof opening in one quick continuous motion, shoulders and back entering first as she draws her legs inside. Follow her through the opening onto the solid opaque interior floor. Preserve the overturned cabin’s orientation.

25–28s: Inside, she braces against the continuing jolts. Snow packs against the exterior of the intact side windows, turning their view white. Interior air remains clear, with her clothing, seats and solid floor visible in dim diffuse light. The torn-off hatch lid remains absent; the entry opening stays open and unobstructed.

28–30s: Continue the reference’s leftward camera move, framing the open roof-hatch aperture. Snow-covered terrain and trees pass outside, revealing that the cabin continues downhill, pushed by the avalanche from behind. Snow stays packed against the side windows while the downhill-facing opening remains clear. She watches the passing terrain through the opening, breathing hard, still shaken and alert. Her expression shows tentative relief while she remains braced for another jolt. Keep her gaze directed toward her surroundings, with no smile or look into the camera.
Maintain the reference’s continuous camera path, natural motion blur, believable weight and physical contact. No added cuts, slow motion or decorative camera moves.
Sound: wind, metal strain, cable snap, powder landing, cabin impact, the pole striking and tearing off the hatch lid, avalanche roar becoming muffled inside the cabin, and her ragged breathing. No music, dialogue or text.