Create million-view miniature art videos using Google Flow and ChatGPT with 3 Prompts

A step-by-step process for producing miniature landscape painting (ASMR) videos and bringing million-view paintings to life using ChatGPT and Google Flow Omni Flash with just 3 master prompts.

The niche of miniature book painting and living art (ASMR) videos is attracting millions of views on TikTok, YouTube Shorts, and Instagram Reels. The magical visual experience begins with meticulous brushstrokes sketching scenic landscapes onto tiny notebook pages, then the entire picture gradually comes to life as a real miniature world, accompanied by the rustling of the brush and the soothing sounds of nature.

In this article, you will learn the complete 4-step process of creating hundreds of vibrant miniature painting videos without any painting skills, using only two free tools: • ChatGPT: Acts as a painting expert, analyzing landmark images and generating the precise painting order from the sky to the foreground details. • Google Flow: A next-generation AI video engine with the Omni Flash model, creating realistic water brushstroke animation (Part 1) and bringing the painting to life from the final frame (Part 2).

Sample concept: "A Miniature Mount Fuji Comes to Life in Your Notebook"

• Setting: An artistic wooden table, a tiny leather-bound notebook (small enough to fit in the palm of your hand) and a meticulously crafted soft-bristled paintbrush. • Part 1 (Painting Scene): The hand holding the brush glides smoothly, painting layers of watercolor: turquoise sky → snow-capped Mount Fuji → calm lake → cherry blossoms in bloom in the corner of the painting. • Part 2 (Living Scene): The static painting on the paper begins to come alive: clouds drift lazily over the mountain peak, ripples gently on the lake's surface, and cherry blossoms swaying softly in the breeze. • ASMR Sound: The sound of the paintbrush lightly scraping against sandpaper, the sound of the brush dipping into water, and the gentle rustling of the mountain wind. • Format: Vertical video, 9:16, consisting of two consecutive scenes (total duration 15–20 seconds).

4-Step Process for Producing Miniature Art Videos with 3 Prompts

Step 1: Master Prompt 1 — Analyze the painting order on ChatGPT

Open ChatGPT, upload the image of the landmark you want to create (e.g., Mount Fuji), and paste Master Prompt 1 below to extract the sequential drawing order:

Analyze this landscape image and break down the realistic watercolor painting sequence into 4-5 chronological stages, from background to foreground.

Output only a single concise descriptive string of the painting order that will be used inside an AI video prompt, following this structure:
"First painting the [sky and distant elements], then detailing the [midground mountains/structures], followed by the [water/ground reflections], and finally painting the delicate [foreground foliage/cherry blossoms] in sharp details."

The ChatGPT output will be similar to: 'First painting the soft pastel blue sky and glowing clouds, then detailing the iconic snow-capped peak of Mount Fuji, followed by the deep blue lake reflections, and finally painting delicate pink cherry blossom branches in the foreground.'

Step 2: Master Prompt 2 — Miniature Brushstroke Animation on Google Flow (Part 1)

Access Google Flow > Create New Project > Project Configuration: • Mode: Select Ingredients and upload the original landmark image to the reference box. • Aspect Ratio: 9:16 (Vertical Video). • Model: Omni Flash (smooth motion, optimized materials). • Duration: 10 seconds (costs 15 credits). • Turn off Watermark in settings.

Paste Master Prompt 2 into the input field, replacing the [LOCATION] and [PAINTING_ORDER] you just got from ChatGPT:

Vertical 9:16 macro cinematic video. A tiny miniature art notebook resting on a rustic wooden desk. A delicate miniature watercolor brush held by gentle fingers is actively painting an exquisite, hyper-detailed painting of [LOCATION: Mount Fuji with cherry blossoms] based on the reference image onto the miniature page. The painting comes together stroke by stroke in this exact sequence: [PAINTING_ORDER: First painting the soft pastel blue sky and glowing clouds, then detailing the snow-capped peak of Mount Fuji, followed by the deep blue lake reflections, and finally painting delicate pink cherry blossom branches in the foreground]. ASMR realistic satisfying wet brush bristles sweeping across textured watercolor paper, subtle brush dipping sound, warm natural lighting, ultra-sharp 8k macro details, smooth 24fps motion, no text, no watermark.

Step 3: Save Last Frame & Master Prompt 3 — Bringing the Picture to Life (Part 2)

After rendering Part 1, perform the technique of transforming the drawing into a moving world:

1. Pause Part 1 video at the last second > Hover over the video and click the 'Save Frame' icon > This frame will automatically be saved to the Uploads tab. 2. Switch the mode from Ingredients to Frames > Select the saved frame as the Start Frame. 3. Paste Master Prompt 3 below and press Generate with Omni Flash Model (duration 10 seconds):

Vertical 9:16 seamless macro camera. Starting exactly from the reference frame of the completed miniature watercolor painting inside the open notebook. The static watercolor painting magically comes alive with gentle, realistic living motion: soft fluffy clouds drift slowly behind Mount Fuji, realistic ripples gently lap across the lake water surface, and the delicate pink cherry blossom petals sway in a soothing mountain breeze. Warm peaceful lighting, crisp natural textures, relaxing ambient mountain wind and distant nature sounds, buttery smooth 24fps living painting effect, no sudden camera shake, no watermark.

Step 4: Join the clips and publish directly in Google Flow or CapCut.

You can combine two scenes into a seamless video directly on Google Flow: • In Google Flow: Open Part 1 clip > Click the (+) 'Add clip' > Select Part 2 clip > Click 'Add to scene'. The two clips will be synchronized and seamlessly joined into a complete video without any additional editing steps. • Advanced CapCut adjustments (Optional): If you want to enhance the visual appeal, you can increase the Saturation to +15 and add the sound of a gentle stream or birdsong at -20dB before exporting the Full HD 1080p video.

Quality control checklist before publishing.

Lean manufacturing system

Ảnh địa danh gốc (Núi Phú Sĩ / Hồ Peyto / Tháp Eiffel...)
→ Master Prompt 1: ChatGPT phân tích thứ tự vẽ cọ (Painting Order)
→ Master Prompt 2: Google Flow render cảnh vẽ cọ thu nhỏ (Ingredients Mode, 10s)
→ Trích xuất Last Frame của Part 1 làm Start Frame
→ Master Prompt 3: Google Flow render cảnh tranh chuyển động sống động (Frames Mode, 10s)
→ Ghép 2 clip trực tiếp trên Google Flow hoặc CapCut
→ Xuất bản video Miniature Living Art triệu view

With just 3 standard prompts combining ChatGPT and Google Flow Omni Flash, you can easily transform any landmark in the world into vibrant miniature works of art, building a YouTube Shorts/TikTok channel with millions of views and a unique, relaxing style.