ResourceSpace and OpenAI integration expands to include GPT-4 image processing

Back in January we unleashed the power of GPT AI for metadata processing via the ResourceSpace OpenAI GPT plugin. With the plugin, users can set prompts in order to automatically generate an output field's value, for example, descriptions, captions, or sets of keywords, from text in an input field.

Our OpenAI integration has now been extended to include GPT-4 multimodal support which means that the source for this output can now be the image itself. With this added capability, it’s now possible to tell GPT what to produce directly from an image during the upload process. 

To view this video please enable JavaScript, and consider upgrading to a web browser that supports HTML5 video

OpenAI GPT-4 multimodal integration will save users a lot of time when it comes to adding metadata, keywording, and writing lengthy captions. 

Here's an example:

GPT4 Sample Image

Setting a prompt within the ‘Image description’ field with the OpenAI plugin enabled

GPT4 Prompt

The resulting description

GPT4 Description

Here's the text:

The image depicts a cityscape during what appears to be either sunrise or sunset, given the warm golden hues in the sky. The foreground is dominated by a series of rooftops, featuring various architectural details such as chimneys, dormer windows, and metallic structures, possibly elements of roof terraces. In the background, two iconic structures are visible: the Eiffel Tower, which stands prominently on the left side of the image, and a Ferris wheel to the right, both silhouetted against the bright sky. The overall atmosphere of the image is serene and picturesque, capturing the romantic essence often associated with the cityscape it represents.

The prompts are entirely user configurable, so there are virtually no limits on the potential applications this has when it comes to metadata processing. At the moment OpenAI has imposed a limit of 100 requests per day for their preview model, but this will be removed once the functionality is rolled out in their standard GPT-4 model.

We’re excited to bring this feature to ResourceSpace - giving users even more options and freeing up valuable time that can be better spent overseeing content rather than manually inputting lengthy metadata.

Ready to learn more?

Are you interested in finding out more about how ResourceSpace will benefit your organisation? Our DAM experts are always happy to help.

Alternatively, you can launch your free ResourceSpace instance within minutes. Simply complete the sign-up form on the link below and you're ready to go. It's completely free with no usage limits, so what have you got to lose?