小白研究所

Create a continuous anime video generation workflow. 
Inputs: 1. A file upload named 'Previous Video'. 2. A text input named 'New Scene Script'.

Process 1: Add a multimodal LLM node. It must analyze the 'Previous Video' to extract character designs and art style, then combine it with the 'New Scene Script' to write a highly detailed, visually consistent prompt for the next scene.

Process 2: Add a Video Generation node. It takes the text prompt from Process 1 and uses the 'Previous Video' as a visual reference (Image/Video to Video) to generate an 8-second anime-style video.

Output: Display and output the generated video file.

You are a master anime storyboard director and visual prompt engineer.
Your task is to maintain absolute visual consistency between scenes.
Analyze the Reference: Carefully watch the uploaded previous_video. Extract the specific anime art style, character details (hair color, clothing, accessories), and the final lighting/environment of the last frame.
Integrate the New Action: Read the new_scene_script.
Generate the Prompt: Write a highly descriptive, comma-separated English prompt for an AI video generator.
Format Rule: Your output must ONLY be the final English video prompt, nothing else. No conversational text.
Example Output: High quality anime style, a red-haired swordsman in a black cloak, standing in a ruined city, dark clouds, holding a glowing sword, he swings the sword forward creating a massive fire wave, dynamic action, cinematic lighting, masterpiece.

An epic masterpiece anime style scene. A lone samurai, with sharp crimson (red) hair and a deep blue traditional haori, stands in a desolate bamboo forest at sunset. He is holding a glowing sword. He is ready for battle. Cinematic lighting, masterpiece.

Masterpiece anime style, precise composition. As an immediate continuation of the previous scene, the crimson-red haired samurai, still wearing his deep blue haori in the same sunset bamboo forest, grips his glowing katana. He executes a devastating downward slash. A colossal, roaring wave of intense realistic fire erupts from the blade, scorching the bamboo. Dynamic camera angles, volumetric smoke, high quality, cinematic lighting, cinematic shading.

Build an AI app for creating continuous 3D Q-version (Chibi) viral meme animations.
Inputs:
A Video upload field named "Previous Meme Video".
A Text input field named "Viral Meme Topic (Chinese)".
Workflow Steps:
LLM Concept Artist Node: Pass both inputs to a Gemini model. Instruct it to:
Analyze the "Previous Meme Video" to extract the specific 3D Chibi character design (big head, expressive eyes, cute proportions) and the vibrant, toy-like art style.
Translate the Chinese "Viral Meme Topic" into a funny, exaggerated English animation prompt.
The LLM MUST automatically prepend these rules: "High-quality 3D Chibi animation, cute Q-version aesthetic, vibrant 3D lighting, smooth squash-and-stretch motion, Vertical 9:16 aspect ratio, full body centered."
Video Generation Node: Pass the enhanced English prompt to a Video model (like Veo). Configure it to output in a 9:16 vertical format.
Output:
Display the resulting continuous Q-version animation.

A highly detailed, top-tier 3D Chibi (Q-version) anime character with a huge head and small body. The character has massive, expressive, cute eyes and is wearing an oversized bright yellow hoodie with chunky trendy sneakers. The art style is premium 3D designer toy (like Pop Mart blind box figures) with octane render quality, showing highly realistic material textures, glossy vinyl elements, and vibrant, cinematic lighting. Set in a clean, bright, pastel-colored minimalist studio. The Chibi character is directly facing the camera, performing a viral, dynamic, and exaggerated TikTok-style meme dance with smooth squash-and-stretch animation (e.g., bouncy hip sways and sassy head tosses). Vertical 9:16 aspect ratio smartphone video. EXTREME WIDE SHOT. FULL BODY SHOT from head to toe. The entire character, including the top of the head and the chunky shoes, MUST be fully visible and perfectly centered in the frame. Do not crop.

Vertical 9:16 aspect ratio smartphone video. EXTREME WIDE SHOT. FULL BODY SHOT from head to toe. The entire character MUST remain perfectly centered within the frame without any cropping of the head or the chunky shoes during large movements. Seamless continuation. The exact same highly detailed 3D Chibi character, wearing the identical oversized bright yellow hoodie and chunky sneakers. The character effortlessly performs the viral "Slick Back" floating glide dance step. The animation features incredibly smooth, exaggerated squash-and-stretch cartoon physics. The character's facial expression dynamically transitions from cute to a highly confident, sassy smirk. Maintain the exact premium 3D designer toy aesthetic (Pop Mart blind box style), glossy vinyl material textures, and the identical clean, bright pastel-colored minimalist studio background. High quality, octane render.

Vertical 9:16 aspect ratio smartphone video. EXTREME WIDE SHOT. FULL BODY SHOT from head to toe. The entire figure MUST remain perfectly centered within the frame without any cropping of the head or shoes at any point. Seamless continuation from the previous sequence. The exact same highly detailed 3D designer vinyl art toy mascot (Pop Mart style collectible). It is a stylized trendy adult mascot figure, completely CGI and made of high-quality plastic/vinyl, wearing the identical oversized bright yellow hoodie and chunky sneakers. The vinyl mascot performs a high-energy, rhythmic shuffle dance with incredibly smooth and snappy 3D animation. The sequence concludes with the character striking a confident, cool peace sign pose directly at the camera with a sassy smirk. Maintain the exact premium glossy vinyl material textures, vibrant cinematic lighting, and the identical clean, bright pastel-colored minimalist studio background. High-end octane render, fluid movement.

Build an automated AI app for generating highly-ranked SEO blog articles and matching cinematic cover images.
Inputs:
A Text input field named "Target SEO Keyword / Topic".
Workflow Steps:
"SEO Master Copywriter" (LLM Node): Take the input keyword and write a comprehensive, highly engaging, professional SEO blog post in Chinese. The output MUST use Markdown formatting with clear H2 and H3 headings, bullet points, and an engaging introduction and conclusion. Focus on high keyword density and readability.
"Thumbnail Prompt Engineer" (LLM Node): Analyze the generated Chinese article to extract the core technological theme. Automatically write a highly detailed, professional English prompt for generating a blog thumbnail. The prompt MUST include: "Horizontal 16:9 aspect ratio, highly detailed, futuristic tech aesthetic, UI/UX conceptual design, cinematic lighting, masterpiece."
"Blog Cover Generator" (Image Generation Node): Pass the English prompt from the previous step to an image model. Configure it to generate a high-quality blog cover image. Set the aspect ratio strictly to 16:9.
Output:
Display the final Markdown SEO article text alongside the generated 16:9 cover image.