Working with Images
What neural networks can do and how to pose a problem.
Generating images is the second most popular task after text. Midjourney, Kandinsky, Flux, DALL-E - there are plenty to choose from. But what’s more important is not the tool, but the ability to describe the result in words so that the neural network understands.
What to choose - tools 2026
| Tool | Strength | Access to the Russian Federation | Price |
|---|---|---|---|
| Flux Pro | Photorealism, anatomy | Via VPN | Freemium |
| Midjourney | Creative, style, art | Via VPN | from $10/month |
| Kandinsky (Sberbank) | Russian language, no VPN | Without VPN | Free |
| Nano Banana Pro | Creatives with text | Without VPN | Freemium |
| DALL-E 3 | Understanding complex queries | Via VPN | In ChatGPT Plus |
Prompt structure for an image
[Subject/object] + [action] + [environment] + [style/lighting] + [format/composition] + [technical parameters]
Example: “Marketer at a laptop in a dark room, the light from the monitor falls on the face, cyberpunk aesthetics, neon accents, cinematic lighting, 4:3 format, high detail” does not work without details - add specifics: brands on the screen, the city outside the window, time of day.
Build the prompt step by step
Fill in the fields - the constructor will assemble the prompt according to the correct structure. Copy and paste into Kandinsky, Midjourney or Flux.
Try it yourself
Now describe the task in words - AI will turn it into a detailed prompt for generation.
AI answers are free, but limited by the overall site limit. If the limit is reached, come back later: repeated clicks will not speed up the response.
Generate a prompt for a neural network to create images based on the description: [describe the idea]. The prompt must contain: 1. Subject (who/what is in the center) 2. Environment (background, location) 3. Lighting and color scheme 4. Style (photorealism, art, minimalism, cyberpunk) 5. Format (16:9, 4:3, square) 6. Technical details (high detail, 8k, cinematic) Give me a ready-made prompt in Russian.
Which tool to choose for generation without VPN and for free?
What is most important in a prompt for an image?
Is it possible to publish AI creative in advertising without verification?