OpenAI’s text-to-image generator that started making it to the headlines a few months ago, DALL-E can now be directly integrated into applications and products with the release of an API. The company announced on Thursday that this will enable developers to integrate the company’s text-to-image generation technology into their apps. The company also mentioned that more than three million people will be able to use DALL-E to produce over four million generations on a daily basis.
To enable the API, you need to register with Open AI and you require a private API key that will provide access to the AI generator. The API will give three methods for image interaction: Generating images from the scratch, editing an existing image using new text prompts, and creating variations of existing images. The generations are available in three size ranges i.e. 256×256, 512×512, and 1024×1024 pixels with the smaller sizes being faster to generate. The pricing for the generations will also depend on the resolution. The generation of one 256×256, 512×512, and 1024×1024 pixels will cost $0.016, $0.018, and $0.02 respectively. Organizations working with OpenAI’s enterprise will enjoy volume discounts.
Luke Miller, a product manager that is working on the API at OpenAI, told The Verge, “We already have a few customers building on this in very interesting ways.” He also mentioned that some of these are creative while others a more business oriented and he also gave the example of Mixtiles, a startup using the API to create art and posters for home decor, and CALA, which is using the API to help customers design clothing for themselves.
he also gave the example of Mixtiles, a startup using the API to create art and posters for home decor, and CALA, which is using the API to help customers design clothing for themselves.
Open AI previously announced its association with Microsoft for its Image Creator tool available on Bing and Microsoft Edge and its new Designer App. Shutterstock also collaborated with DALL-E, allowing its users to generate images for prompts that are not a part of their library.
With companies collaborating with the AI tool and the release of its API, concerns revolving around AI text-to-image generators remain intact. Although the company mentioned that it has enabled filters that will restrict users from creating images that will be harmful to society, there are also a lot of legalities involving these generations. While DALL-E was leading the pack around nine months ago, there are many other image generators like Stable Diffusion and Midjourney with fewer restrictions that have come into the picture with the same concerns.