Stable Diffusion
Stable Diffusion is a neural network that generates images from a text prompt.
Welcome! Today, we’ll journey from the simplest generations to the most elegant ones, and I’ll provide examples along the way.
Before we begin, I’d like to mention that AI models are excellent at generating various types of content: girls, cities, guys, nature. However, we’ll focus on cats and something unusual to showcase the full potential! You can experiment on your own after reading this guide 😃
I recommend three bots for use, but I will only explain the first one in detail, as the other two are quite similar:
First Steps
Let’s start with the basics. We’ve just entered the bot and want to generate, say, a frog in a pink spacesuit. So, we’ll write:
You’ll notice that it doesn’t matter whether you use Russian or English for the bot, so for clarity, I’ll use English in the following examples.
Upscale Function
Want to improve the quality of your image? Click the Upscale button under the art, and you’ll get a high-resolution PNG image.
Main Bot Functionality
Access the settings via the /menu command or by clicking the ⚙️ Menu button available under each generation.
Format, Model & Styles
📐 Format – The aspect ratio of your generation. The image can be horizontal (for landscapes), vertical (for portraits), or square.
📺 Model – The foundation of your generation, determining the style. All models lean toward realism but retain their unique flair. Try them out and find your favorite! ❤️
🎨 Style – Easily customize your generation. Be sure to try each one. Examples of all styles are available in the bot!
Models
The bot includes 4 different models: SD 1.5, SDXL 1, SDXL Turbo, SDXL Turbo X2.
You get 8 free generations every day.
How to Earn Generations?
We provide you with several free generations daily. If you run out, here’s how to get more:
Referral System
In the ⚙️ Menu, select ✨ Invite a Friend and get your unique referral link in this format:https://t.me/StableDiffusion_robot/?start=
For every person who clicks your link and creates art, you’ll earn 10 generations!
Shop
Find 🚀 Shop in the menu and purchase additional generations.
Developing and maintaining each model costs us a lot, so we’d appreciate your support.
💎 Premium subscription unlocks 18+ art generation.
Generation Contest
When you upscale a generation, a 🏆 Submit to Contest button appears
- Click it—your masterpiece will be sent to our admin,
- who periodically selects works for publication in our channel!
- If your work is chosen, it will be published and you'll receive 30 bonus generations.
Advanced Input: Weighted Tags
Let’s learn how to adjust the weights of tag-words.
a cat, (jumping:0.5), (flowers at the background:1.5), close up
a cat, (jumping:1.5), (flowers at the background:0.5), close up
You can see how adjusting weights changes emphasis.
By default, weight is 1.0. Use values between 0.5 and 1.6 to avoid distortion.
Only round brackets support manual weights
Square brackets just reduce weight by ~10 present
(flowers in the background)→ 1.1 weight((flowers in the background))→ ~1.21 weight- You can also do:
(flowers in the background:1.4) - To decrease, use:
[flowers in the background]or(flowers:0.7)
Characters
We’ve compiled a huge list of character descriptions.
You can find the tags you need or just write the character’s name in Russian—the bot will suggest tags!
Interesting Tags
Non-obvious but powerful tags you can use:
Framing Tips
- Wide shot
- Low angle
- Selfie
- Dynamic pose
- Close up
- Medium close up
- Full body
- Cinematic frame
Tags for Different Genres
The Different Genres
Abstract Art: vibrant colors, geometric shapes, abstract patterns
Surrealism: dreamlike, mystical creatures, twisted reality
Landscape: majestic mountains, lush forests, desert dunes
Portrait: emotive eyes, intense gazes, expressive gestures
Minimalism: simplicity, clean lines, negative space
Realism: hyper-realistic textures, precise details
Pop Art: bold colors, famous faces
Street Photography: candid moments, urban stories
Night Photography: starry skies, moonlit streets, long exposures
Camera Lenses (for prompt enhancement)
Sigma 150–600mm: Great for wildlife/sports.
Example: mountain lion in the wild, Sigma 150-600mm lens, daylight, photoGoPro Max: Ultra-wide shots for vlogging/action.
Example: player kicking soccer ball, GoPro Max, stadium lightsTamron 10–24mm: For landscapes/architecture.
Example: luxury living room, Tamron lens, daylightZeiss Otus 55mm: Sharp, artistic portraits.
Example: classic car, Zeiss Otus, street viewCanon EF 100mm Macro: Insects, flowers, soft portraits.
Example: bee on flower, Canon EF 100mm, daylight
Let see more cool & Contextual Tags
- holographic
- in the style of kawaii aesthetic
- neon light
- pastel accent colors
- surreal fashion photography
- celebration of rural life
- hyperrealistic animal portraits
- liquid foil gradient
- forestpunk
- roman art and architecture
- carpetpunk
- euphoria
- product photography
- algeapunk
- kodak, muted tones, ektachrome
- conceptual installation art
Inspiration
Coming up with a good prompt isn’t easy. You can explore amazing works by our users for ideas.
A photorealistic image of an anthropomorphic wolf with human-like hands and feet,
wearing a suit and tie, playing electric guitar in a bustling NYC street.
Gibson Les Paul, blues riff, dramatic lighting, 8k masterpiece.
Model: RealVis Vanilla
Yes, its a minimal sheet of Totoro
(totoro:1.5), (by hayao miyazaki:2), (minimalism:1.2), (miyazaki style:1.7)
Model: Stable Diffusion XL Midjourney
A simple girl, just girl, right?
17 years old girl, tall growth, silver, thick, long and disheveled hair, light blue eyes, a predatory look, triangular eyebrows, a daring smile, a small chest, a free set.
Model: DreamShaperXL Midjourney
Yes, as you can see, it's a Rick and Morty picture.
Rick from Rick and Morty
Model: Dreamshaper Vanilla
Beautiful nature with an even more beautiful car
white modern car in distance, far far away from camera in fiield with gray green grass.
Model: SDXL Oil painting
She's realy attractive
Skinny model wearing a dress at beach bar, drinking cocktail.
Model: EpicRealism Vanilla
Need more inspiration?
Be my guest: Look Here