About this skill

Problem

Writing “cinematic” or “film-grade” alone often yields generic, plastic, or parameter-inconsistent results. Jimeng AI is sensitive to linear prompt structure, keyword weighting, style anchoring, and negative-prompt limits. This skill turns short creative briefs into executable generation prompts for images and video: main prompt, negative prompt, camera parameters, motion direction, and style constraints.

How It Works

Scene routing: classifies pure landscape, human, animal, or ambiguous input, then applies the matching rule set.
Structure: orders the prompt as subject → action → environment → style → parameters to reduce attention drift.
Parameters: emits defaults such as Arri Alexa 35, 50mm, f/2.0, shutter, ISO, stock, color grade, resolution, and quality terms; human scenes enforce rim or backlight.
Weighting: uses parentheses and keyword:0.7-1.5 values to emphasize key elements while avoiding deep nesting and repeated terms.
Negative prompts: outputs platform-safe base negative terms, with optional additions for people, landscapes, night, or low-light scenes.
Camera moves: supports terms such as dolly in, tracking shot, crane shot, and drone shot, plus start/end frames, lighting shifts, and focus changes.

Boundaries

Best for cinematic stills, storyboard sets, character look design, prop close-ups, and motion-shot scripts. It is not a replacement for full video editing, character-consistency training, or complex 3D asset generation. If the input mixes conflicting styles, such as cyberpunk + ink wash, the skill prefers one dominant style and ties it to concrete parameters.

Screenshots

Use Cases

Turn one-line scene descriptions into executable Jimeng AI main and negative prompts for short-film storyboards.
Specify rim light, shot scale, focal length, costume, and props for character posters while suppressing plastic skin.
Add dolly in, crane shot, lighting shifts, and focus changes to make video-generation prompts concrete.
Match night or low-light shots with 5219 stock, neon backlight, low ISO, and noise-suppression negative terms.

Best For

Directors' assistants creating cinematic concept storyboards who need executable Jimeng AI prompts.
E-commerce visual designers outsourcing look design and needing cinematic prompts for characters, props, and scenes.
AIGC short-video creators building motion shots who need control over camera moves, start/end frames, and emphasis.
Generative AI engineers debugging Jimeng AI prompt structure, weighting, and scene consistency.