Introducing smarter keyword and title generation with ChatGPT!

We’re excited to announce a new feature that will make tagging and captioning your images easier than ever!

Last year, we introduced our AI metadata generation tool, powered by AI engines from Microsoft and Amazon. While this tool has proven effective, the keywords generated by these engines are often generic. This can be useful for example for generalist stock photography sites (or to identify bib numbers, a common use case), but for those with specialized image collections, more tailored tagging is important.

Recently, OpenAI has made significant advancements by adding image analysis capabilities to ChatGPT. This new feature allows for a deeper understanding of images, enabling the model to generate more specific and specialized keywords.

“A cyclist poses at the Col de la Croix de Fer sign, surrounded by stunning mountain scenery and clear blue skies.”
Keywords: adrenaline, adult, adventure, alpine, altitude 2067m, autumn, bicycle, biking, black bike, blue sky, brown grass, Col de la Croix de Fer, cycling, cyclist, d926, direction signs, exercise, exploration, fitness, France, helmet, hiking, La Maurienne, landscape, man, mountain pass, mountains, nature, outdoor, recreation, road, road bike, rocks, scenery, scenic, sign, smiling, sport, St Sorlin d'Arves, sunglasses, tourism, trail, travel, views, yellow jersey

(Generated with the default PhotoDeck configuration. Note how the AI was able to accurately complete the partially-offuscated location and altitude signs.)

We are pleased to announce that this powerful functionality is now integrated into PhotoDeck’s AI metadata generation tool!

A major strength of ChatGPT is its ability to be steered with instructions and context (the “prompt”). And at PhotoDeck, we believe in giving you control over your image management: with this new integration, users can customize the instructions provided to ChatGPT for image analysis.

This means you can guide the model to focus on the specific aspects of your images that matter most to you and use words that are most relevant to your image collection, resulting in more useful, domain-specific keywords.

“The striking architecture of the Oodi Library in Helsinki, Finland, featuring a sweeping wooden facade and large glass windows, creates a modern public space that invites community interaction.”
Keywords: architectural innovation, community center, cultural landmark, curved roof, exterior structure, finland, glass windows, helsinki, landscape integration, modern architecture, oodi library, public space, sustainable architecture, urban design, wooden facade

(We asked the AI to focus on the architectural aspects and materials, and to identify buildings)

Additionally, PhotoDeck allows for the creation of multiple profiles for AI analysis, making it easy to tailor the keyword generation process for different types of images. This flexibility makes our solution not only powerful but also adaptable to your unique needs.

The field of AI is evolving incredibly rapidly, and we are happy to be responsive to these developments when they can bring tangible benefits to our users. We believe that this new feature will be a game-changer for many of them.

To learn more about how to use our AI metadata tool to generate keywords and captions, check out our documentation page!

This integration with ChatGPT is initially BETA, which means that it can change without notice and that we welcome feedback more than ever!