As a photographer whose heart beats for the analog philosophy and the pure black-and-white aesthetics of the Leica M. I am constantly looking for new ways to capture moments and light. The GPTcam is my latest interactive art and technology project. It forms a fascinating, experimental counterpoint to classic street photography. It is a custom-built kiosk station based on a Raspberry Pi, merging physical hardware, custom Python scripts, and state-of-the-art artificial intelligence (OpenAI). A minimalist, full-screen touch interface allows users to control various experimental recording modes. The output is immediate and intentionally so as a raw, high-contrast thermal print. The GPTcam is more than just a technical gadget; it is an interactive installation that invites the observer to re-experience the creation process of images whether through light or pure algorithms.

 

• Vision-to-Text (CAMERA-GPT): A visual-literary experiment. The camera captures an image of its surroundings and sends it to the GPT Vision AI. The AI interprets the image and writes a matching text (available in German, Italian, and English). The result—both the photo and the AI's interpretation—is printed directly as a receipt-style piece of art. Simultaneously, the work is emailed to my digital archive.

• Voice-to-Image (DALL-E MIC): The acoustic canvas. In this mode, the camera remains blind. Instead, a microphone records spoken words, wishes, or abstract ideas. The speech is sent via the OpenAI API to DALL-E, which generates a completely new image from it, printed by the thermal printer just seconds later.

• The Purist (ONLY CAMERA): A direct homage to the raw, unpolished snapshot. The system acts as a pure photobooth, printing the captured image immediately on thermal paper without any AI intervention.