Lifetime Access - One-Time Payment
No subscriptions - Unlimited generations after upgrade
Don't have a license yet?
Choose your preferred option above to get a license key.
Generator
📚 Library
History
AI Launchers
Upscale
Watermark
📖 Guide
🎯 Target AI Platform Optimizes prompt for selected AI
Midjourney Parameters
Reference Image
📍 Location Settings
Select a country first. Climate and landscape matching will follow the selected location automatically.
Camera
Camera Settings
Camera angle stays locked to the reference by default to avoid prompt conflicts. Only lens selection is exposed here.
🎭Scene Elements
🏗Building Weathering
🛣Road & Ground Weathering
> People
> Vehicles
> Nature & Landscape
> Lighting & Atmosphere
> Amenities & Props
> Animals
Current Instructions (numbered - click x to remove individual ones)
Default (Hyper-Real)
Visual Style
Mood & Atmosphere (sets Time + Weather)
🚫Negative Prompt Builder
Save & reload your full setup
🌿Landscape Mode
Post Production
Reference Image
Drop image or click to upload
Render - Sketch - Floor Plan - Photo
Reference
Type: - Style: - Lens: - Mat: -
Exterior Detected
Generated Prompt AI-ready
Upload a reference image (optional) and click Generate to create your prompt.
Reference
Render
Before
After
Analyzing image...
Scanning
Processing
Building Prompt
Complete
📋 Prompt History
Last 10 generated prompts. Click any to restore it to the prompt box.
No prompts yet - generate your first prompt to see it here.
AI Upscaler & Enhancer Free - HF

Real image enhancement - sharpens details, removes artifacts, improves textures.

Free at huggingface.co/settings/tokens -> New Token
Real-ESRGAN: GAN-based - removes artifacts, sharpens edges, enhances textures. Best for ArchViz renders.
Drop or click to upload
PNG - JPG - WebP - Max 10MB
Live Preview & Result
Enhancing via Hugging Face...
First use: 20-40s (model loading)
Upload an image to see the
live preview and add watermarks.
Powered by Ruby - upscale requests go through SketchUp's Ruby layer, bypassing browser CORS restrictions. No "failed to fetch" errors.
Watermark Tool

Upload images, configure watermark, preview on the right - then download all at once.

Upload Images to Watermark
Click or drag images here - PNG / JPG / WebP
Select multiple files for batch watermarking
Watermark Settings
Text
Both
Pattern
Text Watermark
Watermark Preview
Upload images on the left - First image previews here with watermark - Drag watermark to reposition
Upload images on the left
Preview appears here with watermark
Generated Prompts
No history yet.
Generated prompts will appear here.
AI Image Generators
Click any tool to open it. Your generated prompt is auto-copied to clipboard.
KM AI ArchViz Generator
User Guide - v1.8.5
1
LOGIN
Enter your Full Name, Email, and License Key then click Authenticate.
Don't have a key? Click Free Trial to get a 1-day trial (10 generations) or Request to Buy for full access.
2
UPLOAD REFERENCE IMAGE
Option A - Upload Image: Click Browse or drag a render/sketch/photo image into the drop zone.
Option B - No image: Fill in all settings manually and generate without a reference.
Auto-Scan AI: When enabled, the app reads visible building cues, materials, lens hints, lighting, and scene context, then applies them without breaking reference lock.
Tip: Export a clean screenshot from SketchUp first, then upload it here for best results.
3
CONFIGURE SETTINGS
Core settings: Project Type, Building/Room Type, Country/City, Site Type, Time of Day, Weather, and Target AI Platform.
Advanced Settings: Camera, motion blur, scene elements, weathering, material bias, and landscape controls stay hidden until you switch Advanced ON.
Visual Preset Mode: Uses library/admin preset prompts as direct templates. Country, site, time, weather, and project fields can still be adjusted.
Negative Prompt Builder: Exterior and Interior use mode-specific negatives. Floor Plan uses the built-in floor-plan negative set automatically.
Tip: Click INSPIRE ME to auto-select a strong settings combination when you are not using Visual Preset Mode.
4
GENERATE PROMPT
Click > Generate Prompt. The photo template engine rebuilds the prompt from the current settings, normalizes conflicts, preserves selected values, and formats the final output for the chosen AI platform.
Click Optimize for an optional AI enhancement pass. Click Copy to copy the final prompt.
5
USE IN AI TOOLS
Paste your generated prompt into Gemini, ChatGPT Images, Midjourney, Stable Diffusion, Firefly, Flux, or other supported platforms.
The engine applies platform-specific structure, negative prompt handling, and character-limit behavior automatically.
Other Features
Library Presets
Browse admin-curated presets, apply their settings, copy full prompts, or save favorites into My Library.
Mobile / Tablet
The generator stacks settings, upload, prompt output, and history on small screens with larger touch targets.
Negative Prompts
Exterior and Interior negatives sync with the visible builder. Floor Plan uses a built-in safe negative block.
Visual Presets
Use admin preset templates directly, while still adjusting country, site, weather, time, and project fields.
Trial Plan Limits
- 10 prompt generations per 24 hours
- Watermark is always applied to generated prompts
- Library, AI Launchers, Upscale, and Watermark tabs are locked
- Upgrade to Full for unlimited access
Need help? YouTube Tutorials - Get Full License
© KM Blueprint Studio - v1.8.5
Detailed Function Reference
GENERATOR - Every Setting Explained
Auto-Scan AI
When checked, the AI analyzes your uploaded image for project type, style, materials, lens, lighting, and visible building cues. Applied scan descriptions are inserted directly into the prompt without changing reference geometry.
Project Type
Exterior = outside view. Interior = indoor room. Floor Plan = 2D architectural layout. This controls the entire direction of the generated prompt.
Building Type
Category of the structure - Residential House, Villa, Hotel, etc. Affects material suggestions, landscape context, and stylistic keywords.
Room Type
Visible only for Interior projects. Specifies the exact room - Living Room, Master Bedroom, Kitchen, etc. - for more targeted furniture and lighting prompts.
Ref Type
What kind of image you uploaded: 3D Render, Sketch, Floor Plan, AI Image, or Photo. Heavier realism enhancement is applied to renders and sketches.
Ref Lock
ON = AI strictly preserves original structure, geometry, and proportions. OFF = Looser interpretation with more creative freedom from the AI.
Building Age
Brand New = pristine surfaces. 1-5 Years = slight wear. 50-100 Years = heavy patina. Sets the weathering tone for all materials in the prompt.
Aspect Ratio
Output image dimensions. Match Reference uses your uploaded image proportions. 16:9 for widescreen presentations. 9:16 for social/portrait.
Country / City
Sets location context - local vegetation, street character, architecture style, and auto-detects typical local vehicles and climate.
Time of Day
Controls sun position and sky color - Morning, Golden Hour, Midday, Sunset, Blue Hour, Night. Directly affects all lighting keywords in the prompt.
Weather
Clear, Overcast, Cloudy, Rainy, Foggy, Snowy. Weather changes sky, ground reflections, and atmospheric mood in the final prompt.
Shot Style
Default (Hyper-Real), Cinematic, Dramatic, Golden Hour, Aerial, Wide-Angle. Applies a global mood and lens preset over your other settings.
Camera Lens
Focal length determines field of view - 16mm ultra-wide shows more context, 50mm is natural, 85mm+ compresses depth for dramatic facades.
Material Bias
Check materials to emphasize them in the prompt - Concrete, Wood, Glass, Metal, Stone, etc. Use Custom to add your own (travertine, corten steel, etc.).
Additional Instructions
Free-text field for specific people, vehicles, props, mood, or style notes. Vehicle wording can trigger compact realistic plate details when cars/motorcycles are intentionally requested.
Advanced Settings
Keeps expert controls hidden by default on desktop, tablet, and mobile. Switch ON for camera/reference controls, scene elements, motion blur, weathering, material bias, and landscape options.
Visual Preset Mode
Uses selected Library presets as the prompt template. Preset country, project type, site, time, and weather load automatically, but you may adjust them before generating.
Negative Prompt
Exterior and Interior use their own selectable negative prompt sets plus custom text. Floor Plan hides the builder and uses a built-in plan-safe negative block.
Inspire Me
Auto-selects an optimal combination of settings based on your Project Type and Building Type. Great starting point if you're unsure what to set.
> Generate Prompt
Sends your image and all settings to the AI. Returns a fully structured architectural visualization prompt ready to paste into any AI image generator.
Optimize
After generating, click Optimize to have the AI further enhance and refine your prompt for richer, more detailed output.
LIBRARY - Preset Management
Presets tab
Admin-curated master presets visible to all users. Presets include searchable codes, climate/site tags, full prompts, and apply-ready settings.
My Library tab
Your personally saved prompts. Add any generated prompt via the + Save button. Stored locally on your device.
+ Save button
Located next to the search bar. Saves your current generated prompt with a custom name and category into My Library.
Search
Filters presets by name, category, or group in real-time as you type.
Admin button
Opens the Admin panel for creating, editing, and deleting master presets. Requires the admin master key. In admin mode, a Delete button appears on every preset card.
Copy Prompt
Copies the preset's full prompt text to clipboard, ready to paste into any AI tool.
Apply Settings
Loads the preset's saved settings into the Generator. If Visual Preset Mode is ON, applying a normal preset turns it OFF so full settings load correctly.
Delete
Removes the preset. For My Library items, always visible. For master presets, only visible in Admin Mode.
HISTORY - Generated Prompt Log
Auto-Saved
Every successful prompt generation is automatically saved to History with a timestamp. Up to 50 entries are kept.
Copy
Hover over any history card and click Copy to copy that prompt to clipboard.
Use
Loads the reference image from that history entry back into the Generator as a new reference.
Clear All
Permanently removes all history entries. This cannot be undone.
AI LAUNCHERS - Quick-Open AI Tools
Click any launcher card to auto-copy your prompt and open that AI tool in your browser. Featured tools have the best results for ArchViz:
Midjourney
Best overall quality for ArchViz. Use /imagine command in Discord, paste your prompt.
ChatGPT/DALL-E
OpenAI image generation. Fast, great for iterating quickly. Paste prompt in chat.
Google Flow
Google's ImageFX platform. Excellent for fast architectural mood renders and quick previews.
Leonardo AI
Has dedicated ArchViz model presets. Excellent photorealistic exterior results.
Stable Diffusion
Open-source, runs locally. Full control over model, LoRA, CFG scale.
Adobe Firefly
Commercial-safe AI. Great for high-resolution print-ready outputs.
Ideogram / Krea
Strong for text-in-image renders and real-time AI art exploration.
WATERMARK - Protect Your Work
Upload Images
Upload single or batch images (PNG/JPG/WebP). First image auto-previews on the right immediately. Click thumbnails to switch preview.
Text tab
Add custom text watermark. Set font size, color, opacity, and position.
Image/Logo tab
Upload a PNG logo. Control size (% of image width) and opacity.
Position
Choose corner/center snap position. Or drag the watermark overlay directly in the preview to place it exactly where you want.
Drag to Reposition
Click and drag the watermark text/logo in the right preview panel to position it freely on the image.
Watermark All
Applies watermark to every uploaded image and downloads them all - none are skipped or left pending. Downloads stagger automatically to prevent browser blocking.
UPSCALE - Enhance Resolution
Hugging Face Key
Free at huggingface.co/settings/tokens. Create a Read token, paste it here. No payment needed.
Real-ESRGAN 4x
Best overall - removes CGI artifacts, sharpens edges, enhances all textures. Recommended for ArchViz renders.
Swin2SR 4x
Best for real photographs. Optimized for natural photo degradation patterns.
Batch Mode
Queue multiple images and upscale them all in sequence directly from your browser - no extra software needed.
First use ~30-40s
The Hugging Face model needs to wake up on first use. Subsequent uses are faster (usually 5-15 seconds).
Contact & Support
For bugs, license issues, feature requests, or payment help:
hello@kmblueprintstudio.com
We typically reply within 24-48 hours on business days.

YouTube Tutorials   -   Gumroad Store
© KM Blueprint Studio - v1.8.5 hello@kmblueprintstudio.com YouTube - Store