From model-bridge
Generate optimized prompts for LTX Video and create short video clips locally.
How this skill is triggered — by the user, by Claude, or both
Slash command
/model-bridge:ltx-video-promptThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
Generate optimized prompts for LTX Video and create short video clips locally.
Generate optimized prompts for LTX Video and create short video clips locally.
Use this skill when the user wants to:
Video prompts are fundamentally different from image prompts. You're choreographing motion over time, not describing a static scene.
Think like a director writing shot notes, not a photographer describing a still.
Write 4-8 descriptive sentences using this pattern:
[Scene anchor] + [Subject + action] + [Camera movement] + [Visual style] + [Motion cues]
Six-part shot-note format:
| Term | Meaning |
|---|---|
dolly in/out | Camera moves toward/away from subject |
track / tracking shot | Camera moves laterally alongside subject |
pan left/right | Camera rotates horizontally (pivots) |
tilt up/down | Camera rotates vertically |
crane up/down | Vertical movement emphasizing scale |
orbit | Camera circles around subject |
handheld | Natural shake and movement |
tripod-locked | Static, stable frame |
steadicam | Smooth floating movement |
| Duration | Sentences | Action Level | Best For |
|---|---|---|---|
| 3-6 sec | 2-4 | Simple, single action | Product reveals, reactions |
| 6-12 sec | 4-6 | Moderate motion | Scene establishing, demos |
| 12-20 sec | 6-8 | Layered action | Narratives, cinematic moments |
Speed:
slow motion, time-lapsesmooth continuous motiongradual vs suddenSpecificity (reduces artifacts):
5 degrees per second rotationpush-in 2 metersdolly 3 meters left./scripts/video-generate.sh "YOUR PROMPT HERE"
./scripts/video-generate.sh "YOUR PROMPT" output-name 32
When user requests a video:
Understand the scene - What's the subject? What motion? What mood?
Determine duration - Match complexity to length
Write the prompt:
Generate and review
Refine - Adjust motion, pacing, or camera as needed
A sleek smartphone rests on a matte black surface. The camera performs a slow
dolly-in, starting from a wide shot and gradually moving to a close-up of the
screen. Soft studio lighting creates subtle reflections on the glass surface.
Minimal camera shake, smooth continuous motion.
A red fox emerges from tall golden grass at dawn. The camera tracks alongside
at a low angle as the fox moves through the field. Soft morning mist catches
the warm backlight creating a halo effect around the fur. The fox pauses, ears
perked, then continues walking. Handheld documentary style with subtle natural
shake, shallow depth of field on 85mm lens, warm color grading.
A lighthouse keeper walks along the rocky shore at sunset. The camera tracks
alongside him with smooth lateral movement, keeping him in the left third of
frame. Waves crash against the rocks in the foreground while seabirds circle
overhead. The golden light gradually intensifies as the sun dips toward the
horizon. The keeper pauses, looks out at the sea, then continues walking.
Cinematic 2.39:1 aspect ratio, film grain, warm golden hour color science.
A hawk gliding in slow circles against a clear blue sky, wings fully extended.
The camera tilts upward to follow the bird's descent, holding steady as it
passes overhead. Smooth continuous motion, 5 degrees per second rotation.
Natural lighting, telephoto compression at 200mm equivalent, nature documentary
style, minimal background distraction.
Dust motes float through shafts of light in an abandoned library. The camera
performs a slow dolly forward through the space, passing between towering
bookshelves. Books lie scattered on the floor, pages rustling slightly in an
unseen breeze. The light shifts subtly as clouds pass outside unseen windows.
Tripod-locked smooth motion, wide angle 24mm, desaturated color palette with
warm highlights, atmospheric and contemplative mood.
These specifications reduce common artifacts:
npx claudepluginhub lando-labs/claude-plugins --plugin model-bridgeProvides prompting techniques for AI video generation models on Replicate. Covers scene description, camera direction, and cinematography language for better video outputs.
Generates videos from text prompts via fal.ai models like Kling 2.6 Pro, Sora 2, LTX-2 Pro, Runway Gen-3 Turbo, Luma Dream Machine; supplies endpoints, durations, aspect ratios, prompt structures, TypeScript/Python code.
Generates AI videos using Google Veo models via nano-banana CLI for text-to-video, image animation, lip-sync dialogue, and scene extensions.