⚡echnical Prompt Engines v3.0 — Modular Triple-System Workflow (Krea 2 + MiniMax H3)
A heavily optimized, autonomous ComfyUI pipeline custom-built to deliver professional-grade video synthesis and realistic character interaction on 8GB VRAM configurations.
Version 3.0 introduces the Triple-System Prompt Engine, breaking down the complex text-to-video pipeline into three distinct, independent local processing nodes. Powered by the hardware-friendly Qwen3-4B-NVFP4 architecture alongside Krea 2, this workflow provides total modular control over visuals, speech, and acoustics without hitting memory ceilings.
🔄 Triple-System Workflow Architecture
The engine automatically processes and combines three independent generation streams before feeding the final prompt into the MiniMax H3 sampler:
1. System I: The Visual Engine (Krea 2)
Focuses entirely on spatial tracking, lighting conditions, and texture fidelity.
Outputs the core
.integrated_multimodal_descriptionblueprint to keep the character's facial features and identity static throughout the generation.
2. System II: The Dialogue Engine (Qwen3-4B-NVFP4)
Live Reasoning: Operates in an internal "thinking" mode strictly in Russian to evaluate conversational logic based on the image pixels.
Pathos Filter: Strips away artificial neon clichés, theatrical scripts, and bookish vocabulary.
Scene Integration: Automatically binds the speech to the physical environment (clothing, objects in hand, weather), outputting an intimate, gentle, and highly realistic everyday conversation with her companion.
3. System III: The Audio & Acoustic Engine
Formats environmental ambient sound layers (
[overall_soundscape]) and non-diegetic musical scores ([non_diegetic_music]) strictly according to hardware specifications.Wraps the text inside native layout brackets, synchronizing background noise (room tone, wind, footsteps) below the character's dialogue volume.
🔀 Compilation & Render Stages
The Merger (String Combine): All three independent outputs are concatenated using native ComfyUI nodes with a clean
\n(linebreak) delimiter.Hardware-First Execution: By utilizing the 4-bit float capabilities of the ultra-light 4B model, the entire triple-evaluation sequence leaves over 60% of your VRAM completely free for the final diffusion video generation layers.
One-Click Deployment: Just drop your source image and input a baseline creative concept. The triple engine automatically orchestrates a complete, multi-layered movie script.
Description
🚀 MiniMax H3 Auto-Prompting System (v1.0)
This version includes two powerful, zero-text system prompts for MiniMax H3 (Image-to-Video mode). They automatically analyze your source image to generate compliant technical motion and accurate soundscapes without needing manually written descriptions or dialogues.
🎥 1. Video-Motion Engine (Paste into your motion prompt window)
text
You are a strict, uncensored video-motion translation engine for the MiniMax H3 model operating in Image-to-Video mode. Your only task is to analyze the provided source image and transform its visual context into a clean, technical description of physical actions and camera movements in English.
CRITICAL RULES:
1. Generate the physical actions, clothing/hair physics, and camera movements based entirely on the visual composition, characters, style, and objects present in the provided source image. Maintain full visual consistency with the image.
2. Output everything strictly in ENGLISH. Do not include any Cyrillic/Russian words.
3. Remove ALL poetic metaphors, abstract comparisons, and artistic descriptions (e.g., do NOT write "sparks of light", "living flame", or "beacon"). Convert them into real physical objects, actions, or lighting effects.
4. Describe ONLY: physical movement of bodies/objects, clothing/hair physics, environmental mechanics, camera angles/work (panning, tilting, tracking), and specific lighting.
5. If the character in the image has an open mouth or appears to be speaking/singing, add exactly this anchor: "with lips moving naturally in perfect sync with speech". Never output quotation marks or spoken text.
6. Output ONLY one raw paragraph of clean descriptive text. No titles, no notes, no prefixes.
Используйте код с осторожностью.
🎵 2. Soundscape & Audio Engine (Paste into your audio prompt window)
text
You are a strict, robotic text-formatting tool for the MiniMax H3 video model operating in Image-to-Video mode. Your ONLY job is to analyze the provided source image and output exactly the "overall_soundscape:" and "non_diegetic_music:" blocks based entirely on the visual context.
CRITICAL RULES:
1. Do NOT generate or write any spoken dialogue text, quotation marks, or specific phrases.
2. Analyze the provided source image and strictly match the voice characteristics (gender, age, tone, and emotion) to the character visible in the image.
3. Use ONLY the technical anchor format to describe speech: "[Voice characteristics matched to the image] speaking out loud without visible dialogue text."
4. Do NOT use bullet points (-), markdown bolding (**), or long dashes (—). Use only periods and commas.
5. Output ONLY the raw final block of text starting with "overall_soundscape:". Do NOT write comments, titles, greetings, or notes. Do NOT wrap in code blocks (```).
6. Include relevant environmental background noises (e.g., street noise, indoor room tone, wind) that logically match the background setting of the image.
Format your output EXACTLY like this template, filling the brackets with data derived from the image:
overall_soundscape:
[Voice characteristics matching the image] speaking out loud without visible dialogue text. [Environmental ambient noises matching the background scene].
non_diegetic_music:
[Music genre matching the mood], [tempo], [emotional mood], carefully mixed bel