Fatih Ak's InkyPi Picture Frame Project Enters the AI Era with OpenAI DALL-E Image Generation

Developer Fatih Ak has extended the InkyPi Raspberry Pi-powered desktop ePaper display project with a new feature for those who need a little inspiration of a morning: a generative AI plugin that adds artificial art and, optionally, text to the roster.

"I added AI image and text plugins to the InkyPi project I [wrote] previously," Ak explains of the latest extension to the project, which is one of a number of efforts to create low-power always-on desktop and wall-mount displays using ePaper technology. "It uses the OpenAI API [Application Programming Interface] to generate and fetch images and text content from text prompts using the Dall-E and GPT-4o models."

Ak's Python-based InkyPi project is designed to turn a Raspberry Pi single-board computer one one of Pimoroni's popular Inky Impression electrophoretic displays into a smart desktop accessory, which provides the ability to run through a slideshow of images, display newspaper front pages, and work as a clock — all with a sunlight-readable screen that draws no power unless it's in the process of being updated.

The latest plugins for the project extend its capabilities: as well as being able to run through a slideshow of user-provided images, it can now request the generation of new images from OpenAI's DALL-E vision language model (VLM). "I've always wanted a digital picture frame to showcase AI-generated images," Ak explains. "A frame that could display random daily artwork or create visuals from text prompts on the fly."

Running, in the case of Ak's own implementation, on a Raspberry Pi 2 W and a Pimoroni 7.3" Inky Impression display, the image generation — or text generation, if preferred — does not occur on-device, but is farmed out to OpenAI's servers via API calls. As plugins, though, the new features operate alongside the project's existing functionality and can be disabled if preferred.

The latest version of the InkyPi source code, including the gen AI plugins, is available on the project's GitHub repository under the reciprocal GNU General Public License 3, along with instructions for installing it. More information is available in Ak's Reddit post.