Thrilled to announce the release of the first version of our ONiT Explorer at https://travelogues.app/.

The web application allows AI-assisted search within our corpus of images. For now, this means more than 22,000 images we extracted from our travelogues in English, French, German and Latin language that were printed between 1501 and 1850 and survived in the holdings of the Austrian National Library (ÖNB) in Vienna. You can either search for text or similar images.

Please note that the tool was developed for research purposes and is fine-tuned for the retrieval of representations of ‘nature’ (i.e. flora, fauna, landscapes and maps). This concerns more than 8,700 of the images. Be aware that you are interacting with the raw model, which means we do not filter out biased or negative responses. ONiT Explorer is still under development, and may occasionally generate incorrect or misleading information, or produce offensive or biased content.

The images have been scanned in the course of the Austrian Books Online project – a public private partnership between the ÖNB and Google Books – as well as in-house at the ÖNB and by our project team. For the extraction we used the Illustration Detector tool by the Visual Geometry Group at the University of Oxford. Despite our efforts it was not possible to scan all the leaflets in their unfolded state. We are working on replacing as many as possible in the course of the project. For more details on the constitution of our corpus please consult our publications. Feedback is always welcome.

The official release, including expansions in terms of images and usability, will take place on January 20, 2025. More details will follow soon.