1. Latest news
  2. 🏞 Text-to-Image AI Generator
🏞 Text-to-Image AI Generator
Published: January 29, 2023

🥳 Meet the release, service generative images (text-to-image) from Gerwin AI!

Since today, a section "Image Generation" appeared in the user's personal cabinet (temporarily only for "Gerwin PRO" accounts). After the alpha-test, we plan to open the new section for all users of the service. With the new feature, the user can create an infinite number of unique pictures in different styles, from realistic retro photos to juicy and bright concept art.

How to work with images

  1. Create a new task, enter a simple query about what you want to see in the finished image. For example it could be: "Moscow Landscape", "Supercomputer in Space", a portrait of a person, or any other query. Your request will be processed, and you will get a choice of the most suitable styles.
  2. Choose the format (aspect ratio) of the future image. Three options are currently available: square, horizontal, vertical.
  3. Choose the style you like and press the "Draw" button. The neural network will start to draw according to your request and the chosen style 4 new images.
  4. Then, if you wish, you can scale any or each of the finished images. Scaling will allow you to create the selected picture in high resolution.

Image size

You will see a progress wheel during the generation process. After each image is generated, you will be able to copy or download the finished images in small resolution. In square format, the size of just generated image will be 512x512px. After applying scaling, the image will be available in 1024x1024px and 2048x2048px resolutions. For horizontal and vertical formats, the available resolutions will have different but appropriate values.

Uniqueness of images and usage rights

All the generations are unique, the images are created literally from scratch, the percentage of probability of coincidence with other images on the network on various metrics will not exceed 0.001%. All of the user's generations are available in your personal cabinet in the "Gallery of images" in private mode, with a storage period of not less than 30 days. The rights to use belong to the user, the images can be used for personal and commercial purposes, without the obligatory indication of authorship and links to the service.

Remember that the use of the images must not violate the laws of the countries in which they are used. It should not violate anyone's copyrights (in the case of famous personalities, artists, etc.). The results of the generation should be used responsibly without violating moral and ethical standards and should not be used for "black PR", fake news or other tasks that can cause reputational moral or any other harm to other people. Remember that the responsibility for publishing lies with the user.

Cost of generating images

Your balance is charged for each generation of 1,200 characters. One generation includes 4 images. Characters are not deducted for generations that were not created, were not loaded in the system or got under the content filter. Images are not included in the character scaling program, there is no "disliking" function at the moment. For each scaling of 1 image from your balance 1200 characters will be deducted as well.


The modern development of neural networks allows you to create incredibly beautiful illustrations. However, there are a number of points that are in the process of improvement, so, for the best results, they should be taken into account. Difficulties arise most often in the drawing of hands, in the accurate drawing of complex actions. For example, it will be difficult to get a completely relevant image on the query "Little girl picks up a computer mouse from the floor and passes it to her mother, and in the background Dad strokes the dog's neck. Close interaction between people and animals in the frame (eg "cuddling"), will most often be poorly drawn. But to get the right photorealistic portrait or a conceptual picture of the "programming team," for example, is not difficult.

Test mode

Please note that we are not responsible for the generation quality and stability of the image service. The image generation function works in test mode and can be unavailable indefinitely in case of technical problems.

Enjoy your use,
Gerwin AI team! ❤️