Do you have a site (or social accounts) and want to illustrate your publications with unique royalty-free images? Interested in AI and image generators like Dall-E 2? Here are 10 of the best free (and professional) AI generators to create an image from description (text prompt) . These free and paid artificial intelligence apps can turn a sentence into a realistic painting or image, all royalty-free.
Quick Read: Top 10 Automatic Image-to-Text Generators
- Nightcafe, the 100% free abstract image generator with a daily quota
- Lexica.art, one of the best abstract image generators (free)
- Generated.photos, the generator of realistic photos of people
- Dall-E 2, the benchmark for generating realistic images from text
- Crayion, or Dall-E mini, the 100% free alternative to other visual generators
- Text To Image, a free tool to generate images from text, offered by the Canva platform
- JasperArt, one of the tools of Jasper IA, specialized in art creation
- Bing Image Creator, Bing’s tool, created by Microsoft
- Deep Dream Generator, another tool for art lovers
- Photosonic by Writesonic, the most pro generator with a free version up to 10 images (best value for money)
1 – Nightcafe, the easiest free AI text to image converter
Nightcafe is the epitome of those text-to-image apps that will knock your socks off seeing the kind of mind-blowing creations AI can do.
Type any descriptive phrase in English and Nightcafe will use AI to turn it into a painting (this site is not capable of creating realistic images).
When you create a new image, you can choose from various artistic styles (oil paint, matte, surreal, steampunk, etc.). Next, you have to choose between two AI styles: Artistic and Coherent.
There is a fundamental difference between the two styles. The Artistic variant is best suited for abstract creations, such as depicting buildings in the sky or other imaginative phrases. The Coherent version is more suitable for “realistic” images (although the results remain generally abstract).
Here are concrete examples of achievements obtained with this image generation AI and fairly detailed prompts:
With Nightcafe, you can select aspect ratio, output resolution, and a few other settings, and you’re done. Nightcafe will give you an original artwork based on your settings and text prompt in seconds. All your creations will be saved in your online account.
When you create an account, you receive five free credits, as well as five daily credits. Credits determine the number of changes you can make to a build’s settings.
At the end of the generation, you can download your “work” for free in low resolution (.
Website: https://creator.nightcafe.studio/my-creations
2 – Lexica.art: the best free AI image generator (non-realistic)
Lexica.Art is a free artificial intelligence (AI) powered image search engine and generator developed by Stable Diffusion.
It was designed as a tool to find and study images created by AI.
The interface is simple, consisting of a search bar and image previews.
When a search is performed, the results include Source Text, Seed, and Classifier Free Guide Scale, which can be used to recreate a similar image using Dream Studio, another Stable Diffusion project.
Lexica Art also has a feature that allows users to generate images from text, with a limited number of options compared to Dream Studio.
Users can save their favorite images and use them for personal use in a personal gallery, and can also upload an image to use as a starting point for generating new images.
Overall, Lexica Art is a useful and easy-to-use tool for exploring and creating images with AI.
Website: https://lexica.art/aperture
3 – Generated.photos, the powerful image generator of people who seem real
In the same spirit as the “This Person Does Not Exist” site, the photos of people generated by the Generated.photos site are created from scratch by AI systems .
All images can be used for any purpose without worrying about copyright, distribution rights, infringement claims or royalties.
The photos are very realistic and it is possible to customize many criteria on the avatar you want to create with Generated.photos such as the choice of gender, hair type, hair color or skin color, etc…
Are the photos of people generated by Generated.photos free and royalty-free?
Images can be downloaded for free and used for personal use. For any use of an AI-generated profile photo Generated.photos, the site simply requests a backlink to the source in return. Higher quality images and commercial use licenses are available for individual downloads and API access. Prices start at $19.99 per month for the premium version offering access to larger file sizes (1024 x 1024 pixels).
Website: https://generated.photos/
4 – Dall-E 2, the image generator for professionals (with API access)
DALL-E 2 is a new AI system capable of creating realistic images and artwork from natural language description.
Dall-E 2 is a professional image generator with API access . It is able to produce high quality images from simple texts. It is also very easy to use and generates images quickly and efficiently.
Website: https://openai.com/dall-e-2/
5 – Craiyon (dall e mini), a 100% free (but limited) online image generator
Craiyon, formerly DALL-E mini, is an AI model that can draw images from any text for free (the site just runs ads to fund its servers)!
The only drawback, and not the least important, is that the images generated are of low quality, making them difficult to use on websites (256 x 256 pixel format only).
Here is an example of the result obtained with this prompt “A realistic picture of a computer using AI to generate photos”:
The two founders of this AI are Boris Dayma (who trained the current version of the AI model DALL-E mini) and Pedro Cuenca (who worked on the backend).
To get the best out of DALL-E mini, the founders recommend using specific terms such as “illustration”, “realistic photo”, “high definition”… in addition to descriptions of the type of photo sought.
The model used by Craiyon is called ‘ SLAB mini ‘, more specifically the larger version also known as ‘SLAB mega’ and is trained using Google TRC.
Are the images generated by Craiyon free of rights?
There are two types of licenses, each with its specificities:
- The free commercial license;
- The paid commercial license.
Dall-E mini free commercial license
You may use the Site or the Images for academic or research purposes or to educate or entertain on various social media platforms.
If you are an individual or entity with annual revenue of less than US$1 million, you are permitted to use images generated by the Service for the purpose of financial gain, provided that (1) a 20% royalty is due on revenue attributable to any blockchain transaction (such as the sale of NFTs), and that (2) Craiyon may revoke your right to use the Site or images at any time.
Free commercial use requires that you attribute the images to Craiyon (via a link in particular).
Craiyon Paid Commercial License
If you are an individual or entity with an annual income of more than US$1 million, commercial use of the site or images is subject to the terms of a separate commercial license agreement. To find out more, you can contact them at sales@craiyon.com.
Website: https://www.craiyon.com/ / https://huggingface.co/spaces/dalle-mini/dalle-mini
6 – Text To Image, Canva’s free tool to generate an image from text
Text To Image is another free tool for generating images from text, and it’s offered by the Canva design platform.
To use this tool, you just have to enter some text in Canva’s editor and choose the style of media you want to create. You can also choose from a variety of themes, colors and fonts to personalize your image.
Once you have chosen an image style, all you have to do is enter the description text of the image you want to create, then let the generator do its work.
Once the images are generated, all you have to do is choose the one you like best and then save it to your Canva account or download it directly to your computer.
Canva’s Text To Image tool is very easy to use and can save you a lot of time when creating visual content for your website, social media, or presentations.
This feature is especially useful for people who want to add visual flair to their textual content.
7 – JasperArt, one of the tools of the Jasper AI suite
JasperArt, one of the tools of Jasper IA, specialized in art creation.
This tool lets you create digital art from text, using multiple art styles to meet a variety of needs.
Similar to Jasper IA which generates text, Jasper Art is easy to use : just provide text and the tool offers several images. The many styles offered make it possible to generate ultra-realistic images, cartoon type or in any creative universe.
To benefit from Jasper Art, you must first register on the site. It is possible to register and cancel your subscription within 5 days to pay nothing if you just want to try the tool.
Unlike Jasper IA, there is no free trial version for Jasper Art. You will therefore have to pay to benefit from this service.
8 – Bing Image Creator, Microsoft’s AI image generator
Bing Image Creator, Bing’s tool, created by Microsoft, is another AI-powered image generator. The tool is based on an advanced version of DALL-E, OpenAI’s image generation technology, and allows users to create visual content from a text query.
Among the features of Bing Image Creator is the proposal of four images for each search, with the possibility of choosing the image that best suits the user’s needs . The image can be saved in JPG format with a resolution of 1024 × 1024 pixels and stored in the user’s personal space.
A social media sharing tool is also included in the tool.
In addition to these features, Bing Image Creator offers a “surprise me” feature that generates a random request and suggests a result to explore.
At the time of writing, the image builder is only accessible on the dedicated Bing Image Creator website, but Microsoft is planning to integrate it into its GPT-4-based chatbot and Microsoft Edge at the future.
Good to know: Although the interface is available in French, Bing Image Creator currently only works with requests in English .
Bing Image Creator is available for free on all search engines (but only for users with access to the new Bing).
Good to know: use is limited to 25 credits (each image generation = one credit)
9 – Deep Dream Generator, a perfect tool if you are an art lover
Deep Dream Generator, an ideal AI image generator tool for art lovers!
It is an online tool that uses artificial intelligence to generate artistic images from existing photos or images .
This tool is based on the Deep Dream technique, which consists of using artificial neural networks to create original images from existing images.
To use Deep Dream Generator, you can upload an image of your choice, then select an art style and adjust the settings to your preference.
The art styles offered range from abstract to more realistic styles. The tool then applies filters and transformations to the original image to create a new image with the selected style.
Once the process is complete, you can download and save the generated image to your computer.
Deep Dream Generator is an interesting tool for people who want to create unique and original artistic images from their own photos or images.
10 – Photosonic: Writesonic’s professional royalty-free image generator (to replace paid image banks)
Created by the automatic writing tool Writesonic, Photosonic is an online tool that allows you to create realistic or artistic images from any text description, using a state-of-the-art text-to-image AI model.
The model is based on latent scattering, a process that gradually transforms a random noise image into a coherent image that matches text.
You can control the quality, variety and style of the generated images by adjusting the description and re-running the template.
What are the possible use cases for Photosonic?
You can create images of landscapes, animals, objects, people, scenes, or anything you can imagine, and customize them with various attributes and details.
You can also use Photosonic to enhance or modify existing images by adding text annotations or filters.
Is Photosonic free?
Photosonic is free with up to 10 image generator credits which you can use to create up to 20 images from any text description.
You can also earn extra credits by sharing Photosonic around you by inviting others to join the app.
If you want to unlock unlimited image generation and access more features and settings, you can upgrade to a premium subscription for a low monthly or yearly cost (or buy credits a la carte).
Are the images generated by Photosonic free of rights?
Yes. When you use Photosonic, the images you create become public and therefore royalty free (and free for anyone but you to use).
This means that you and other users may freely reproduce, display, perform, distribute, and create derivative works based on these images, without any royalty or attribution requirements.
Can images generated by Photosonic be used for commercial purposes?
You have all the rights to use the images created by Photosonic from your data for personal or commercial purposes, as long as you respect the rules and the conditions of service of the tool.
What is the quality of the images generated by Photosonic?
They can be downloaded up to 2048 x 2048 pixels, which is more than enough for most web media.
Update 2023
Photosonic stops on April 8, 2023. But in reality, the tool does not really disappear. It is simply the Photosonic platform that is disappearing, but the AI image generation service will remain accessible on the WriteSonic site. The service requires a paid subscription, but ensures superior quality in the generation of images and in the number of creative possibilities.
Website: https://photosonic.writesonic.com/
Answers to frequently asked questions about image generators, digital paintings and photos (by AI)
The use of sites capable of generating images from simple text raises many questions. Here are some of the most frequently asked questions around this topic. If you have others, do not hesitate to ask us in comments so that we can answer them and add them to the list.
What is an online AI image generator?
An online AI image generator is a computer program that is capable of creating high-quality images from digital data . These images can be illustrations, logos or photos.
Generally, it only takes one or more words to be able to create unique images using an AI trained for this. The more detailed the input (prompt) text, the more likely the generated image will look like the expected result.
How do AI-powered image generators work?
Image generators powered by artificial intelligences (AI) are computer programs capable of producing high-quality images from random data. These images are created using a machine learning technique known as ” transfer generation “.
Are image generators free?
Image generators are online tools that allow you to create images from predefined texts or templates. These tools are usually free and easy to use, but professional versions are also often offered when the needs are greater than just generating a few images.
Are the images generated by the AI image generator free of rights?
Most of the images generated by the AI image generator are free of rights. You can use them for commercial or non-commercial purposes. If the site or service used does not authorize you to do so, you will in any case be notified when the generated visual is downloaded.
What is Dall-E?
Dall-E is an open-source image generator created by OpenAI, an artificial intelligence research company. Dall-E is able to generate images from descriptive text, such as “a person playing a guitar” or “a dog chasing a Frisbee”. To date, most of the images generated by Dall-E 1 and Dall-E 2 are not yet able to compete with real photos taken by professionals.
How to use Dall E?
Dalle is a free and open source image generator. It allows to generate images of different sizes, depending on your choice. You can also add effects to your images, like rotation, resizing, etc. To use it for the creation of new images, all you have to do is prepare a descriptive text, as complete as possible, describing the image you would like the model to generate for you.
With Dall-E, is Openai the only company to offer an image generator? What alternatives exist?
No, there are different machine learning models trained to create more or less realistic images.
Among the best known, we can cite for example these 4 alternatives:
- X-LXMERT : Created by AI2 Labs, X-LXMERT is an extension of LXMERT, a visual and linguistic connection transformer. The tool comes with training refinements and enhanced image generation capabilities, rivaling specialized image generation models. X-LXMERT features three key refinements: Discretizing visual representations, using uniform masking with a wide range of masking ratios, and aligning pre-training datasets to the right goals. On their project page, the X-LXMERT research team explained the training as follows: “We use Gibbs sampling to iteratively sample features at different locations in space. Unlike text generation, where the left-to-right order is considered natural, there is no natural order for image generation. »
- RuDALL.E : ruDALL-E takes a short description and generates images based on it. The model includes a wide range of concepts and generates completely new images and objects that did not exist in the real world. OpenAI’s Russian model, ruDALL.E, is trained on ruGPT-3, which was trained on 600 GB of Russian text. The Russian model ruDALL.E has 1.3 billion parameters and a YTTM text tokenizer with a dictionary of 16,000 tokens. It relies on a custom VQGAN model that converts an image into a 32×32 character sequence. There are two running models of the tool, Malevich (XL) trained on 1.3 billion parameters with an image encoder and Kandinsky (XXL) with 12 billion parameters. By running the first model with the same text as DALL.E’s last example, namely “an armchair in the shape of a lawyer”, ruDALL.E included the chair and lawyer combination in the form function.
- GLID-3 : GLID-3 is a combination of OpenAI’s GLIDE, OpenAI’s latent diffusion technique and CLIP. The code is a modified version of Guided Diffusion and is trained on photographic-like images of people. Compared to DALL.E, the GLID-3 result is less capable of producing imaginative images for given prompts.
- CLIP : Before launching DALL-E, the OpenAI research team presented an open-access text-to-image tool called CLIP. The Contrastive Language-Image Pre-training neural network was trained on 400 million image and text pairs. The tool efficiently learns visual concepts from natural language supervision and can be applied to classification by providing the names of visual categories to be recognized. In a paper introducing the model, the OpenAI research team described CLIP’s ability to perform various tasks during pre-training, including object character recognition (OCR), geolocation, shares, etc. CLIP has proven to be very efficient, flexible and more widespread. Also, it is much less expensive because it relies on text-image pair datasets already available on the internet. It can adapt to perform a wider range of visual classification tasks.
Can AIs only generate abstract images?
No, not all artificial intelligences only create abstract images, they can also generate relatively realistic images as long as the initial prompt is well detailed.
Can AI be used to generate images in NFT projects?
Can AI-generated images be used in NFT projects? The answer is yes, but on condition that the images are generated by an AI trained to do so (or, at the very least, that the AI user is sufficiently accustomed to the use of these tools in order to get out of them quality pictures). Of course, to be able to be marketed as Non-Fungible Tokens, the images must have been generated using an AI generating royalty-free visuals.
