How to Create Realistic AI Nature Videos with Google Flow and ChatGPT
Workflow Overview
Learn how to produce breathtaking, photorealistic AI nature and travel Reels using ChatGPT and Google Flow. This step-by-step workflow uses a dedicated creative director master prompt to generate viral scene concepts, high-definition 9:16 vertical base images, and smooth cinematic image-to-video animations.
Tools Used in This Tutorial
Try these tools to replicate the exact results
Advanced AI assistant for scripting, creative ideation, and prompt synthesis.
AI workflow orchestration and video timeline tools.
🛠️ Tools & Resources Used
- ChatGPT (AI Prompt Engineer) - Directs the creative concept phase, generating viral nature scenes alongside tailored image and video prompts.
- Google Flow (AI Image & Video Generator) - Generates high-fidelity vertical 9:16 starting images and transforms them into cinematic, fluid nature videos.
⏱️ Quick Workflow Summary
- Step 1: Run the Nature Reels Master Prompt in ChatGPT to generate 10 scenic concept ideas.
- Step 2: Select your preferred scene number to extract paired Image and Video prompts.
- Step 3: Generate the initial 9:16 photorealistic landscape image in Google Flow.
- Step 4: Switch to video mode, set the generated image as prompt input, and animate the natural motion.
- Step 5: Repeat the workflow across multiple concepts to build a cohesive nature and travel Reels library.
📝 Step-by-Step Tutorial
Step 1: Run the Nature Reels Master Prompt in ChatGPT [Timestamp: 01:04]
Open ChatGPT and paste the complete Master Prompt provided below. ChatGPT assumes the role of an expert AI Nature Reels creative director and will return exactly 10 distinct, highly cinematic nature concepts (such as misty mountain valleys, rainy forest villages, or sunset wheat fields) and halt for your input.
Prompt / Settings:
You are an expert AI Nature Reels creative director and prompt engineer.
I want to create highly realistic, cinematic, peaceful and visually satisfying nature-view Reels using AI image generation and image-to-video generation.
The final videos should feel like real footage captured by a professional camera, NOT like obvious AI-generated content.
STEP 1 — GIVE ME 10 VIDEO IDEAS
When I paste this master prompt, immediately give me exactly 10 unique Nature Reel ideas.
The ideas should be based on visually beautiful scenes such as:
- Field + sunset
- Village + rainy atmosphere
- Windy tree beside a lake
- Massive waterfall in a forest
- Mountain clouds
- Foggy road
- Rainy village
- River flowing through a forest
- Golden wheat field
- Peaceful countryside
- Ocean waves
- Monsoon landscape
- Morning mist
- Mountain valley
- Lonely house in nature
But do NOT simply repeat these examples. Create fresh and visually interesting ideas.
Each idea must have:
1. Number
2. Short catchy title
3. One-line description of the scene
Example:
1. 🌅 Golden Field at Sunset
A vast green field moving gently in the evening wind while the sun slowly disappears behind distant hills.
2. 🌧️ Rainy Village Evening
A peaceful village during light rain, with wet rooftops, misty mountains and trees moving in the wind.
Give me exactly 10 ideas and then STOP.
Do not generate prompts yet.
I will choose an idea by sending only its number.
--------------------------------------------------
STEP 2 — AFTER I SEND THE NUMBER
When I send a number from 1–10, generate ONLY these two things:
IMAGE PROMPT
VIDEO PROMPT
Do not give me any unnecessary explanation, steps, tips, headings like "How to use", or additional content.
--------------------------------------------------
IMAGE PROMPT REQUIREMENTS
Create ONE extremely detailed prompt for generating the selected scene as a photorealistic vertical 9:16 image.
The image must look like real professional nature photography.
Include appropriate details such as:
- realistic natural environment
- realistic weather
- natural lighting
- realistic clouds
- realistic vegetation
- realistic water
- realistic atmospheric depth
- detailed foreground, middle ground and background
- natural colors
- cinematic composition
- realistic shadows
- realistic textures
- subtle imperfections
- professional camera photography
- high dynamic range
- sharp but natural details
- photorealistic appearance
- vertical 9:16 composition
Avoid:
- cartoon
- illustration
- 3D render
- fantasy appearance
- artificial-looking colors
- excessive saturation
- plastic textures
- unrealistic objects
- distorted objects
- text
- logos
- watermark
- people unless the selected scene specifically requires them
The image should be composed specifically for a viral Instagram/Reels vertical frame.
--------------------------------------------------
VIDEO PROMPT REQUIREMENTS
Create ONE image-to-video prompt based directly on the generated image.
IMPORTANT:
Do NOT completely redesign the scene in the video prompt.
The video must preserve the exact environment, objects, composition, weather and overall appearance of the image.
Focus mainly on realistic MOTION and CAMERA MOVEMENT.
Include appropriate movement such as:
- slow cinematic camera push-in
- gentle camera movement
- realistic wind
- moving leaves
- moving grass
- flowing water
- drifting clouds
- falling rain
- moving mist
- subtle waves
- natural environmental movement
Use realistic physics and subtle movement.
The movement should feel like real footage captured by a professional camera.
Avoid:
- fast camera movement
- unrealistic object movement
- warping
- morphing
- melting objects
- flickering
- changing scenery
- changing weather
- artificial motion
- unrealistic physics
- sudden zoom
- camera shake unless specifically appropriate
The final video should feel peaceful, cinematic, immersive and realistic.
VIDEO PROMPT should be suitable for AI image-to-video tools such as Google Flow, Veo, Kling, Runway, Seedance or similar tools.
--------------------------------------------------
OUTPUT FORMAT AFTER I CHOOSE A NUMBER
IMAGE PROMPT:
[Write the complete image-generation prompt here]
VIDEO PROMPT:
[Write the complete image-to-video prompt here]
Nothing else.
--------------------------------------------------
IMPORTANT RULE
Every time I choose an idea, make the prompts highly specific to that exact scene.
Do not give generic prompts.
The Image Prompt should create the starting frame.
The Video Prompt should animate that exact starting frame naturally.
The final result should look like authentic cinematic nature footage rather than an obvious AI video.
Step 2: Select a Concept and Retrieve Paired Prompts [Timestamp: 01:33]
Review the 10 generated scene concepts and reply to ChatGPT with solely the number of your selected concept (e.g., 1). ChatGPT will immediately return two synchronized outputs: an ultra-detailed Image Prompt designed for vertical 9:16 natural composition, and a complementary Image-to-Video Prompt focused strictly on realistic organic motion.
Step 3: Generate the Starting Nature Image in Google Flow [Timestamp: 01:46]
Copy the generated Image Prompt and navigate to Google Flow. Open an image generation canvas, configure the aspect ratio to 9:16 vertical format, paste the prompt into the input field, and click generate to create your high-resolution photorealistic nature frame.
Step 4: Animate into Cinematic Video with Image-to-Video Mode [Timestamp: 02:08]
Switch Google Flow to video mode. Click the three-dot menu on your generated image card and choose Add to prompt to lock the image as your starting reference frame. Configure your preferred video model and 9:16 aspect ratio, copy the Video Prompt from ChatGPT, paste it into the prompt field, and generate the final video clip.
Step 5: Scale Your Nature Video Pipeline [Timestamp: 02:49]
After the video finishes rendering, download the completed nature clip. Return to ChatGPT, select another number from your list of concepts, and repeat the generation pipeline in Google Flow to build a batch of high-retention vertical nature videos suitable for YouTube Shorts, TikTok, and Instagram Reels.
💡 Key Takeaways & Pro Tips
- Preserve Environmental Continuity: Do not write video prompts that alter lighting or introduce new elements; focus strictly on subtle physical motions such as drifting fog, rippling water, rustling leaves, and slow camera push-ins.
- Use 9:16 Native Framing: Always set both image and video generation modes to 9:16 to prevent awkward edge cropping and ensure optimal viewing on mobile platforms.
- Seamless Image-to-Video Ingestion: Using the 'Add to prompt' feature on the existing image ensures Google Flow anchors the starting frame accurately without visual drift or morphing.
📌 Original Source & Attribution
- Source Video: How to Make Realistic AI Nature Videos with Google Flow (Full Tutorial)
- Original URL: https://www.youtube.com/watch?v=8jM6hPYpTw0
- Credit: Nine Tech
