 Excellent. Okay, thank you. Okay, so I'm going to talk about how we use transcribes at the National Library of Norway. And this by describing the workflow from scanned images to full text search. Next please. As many of you know, the National Library of Norway has a vast digitized digital collection. So in 2005, we started the digitization program with the ambition to digitize all our collections. And as of 2021, actually now, and in terms of text material, which I will focus on here, this goal is within reach. So all books, about 600,000 have been digitized, and that is a text corpus of about 33 billion words, which is quite a lot. And if you look at the newspapers, even more, it's about 110 billion words, making it one of the biggest text corpora in the world actually available for usage online. But when looking at our manuscripts, as of until now, we haven't been able to give such numbers because we have only, well, fairly recently at least started to do HDR. So next slide please. This is what our landing page of our digital library looks like nb.no slash search, and it provides full text search and full text view. Provided that you have access to the material. So many very many of our books can be accessed and browsing this page from a Norwegian IP address. And many of the older manuscripts are most of them are in the public domain so there could be browsed worldwide. So the DH lab at the National Library of Norway, which I represent together, for example with my colleagues, Ingrid Bayer and Andre Kvossen who are in present at the conference. We offer services, corporate services based on our digital collection so we do, we have a REST API for which provides concordances, collocations and frequency lists. Next please. So we have, we have 15 years of experience in OCR for printed material, books and newspapers, but not so much or until recently not so much experience with HDR. So, until now, up until now we had this, we have had many manuscripts available in the service, but without a text layer. Transcrevus comes to the rescue, we have experienced with it since 2016 and became a member of the read cope in 2020. Next please. So, the National Library of Norway is a founding member of the IIF consortium, and we make heavy use of that in our digital library so IIF was the best starting point for feeding images to transcribers. Next please. And so the workflow when working with our digital manuscripts is that we upload images of digitized manuscripts using the transcrevus API. The images themselves are stored and on service at National Library made available through our IIF API. And we use simply the IIF manifest in that for that purpose for each document. And then we do document analysis, layout analysis and HDR in transcrevus. And we download, we then download the page, the page XML format and convert it to alt XML using the access of T and feed it back to elastic which is our full text search engine and the NBA National Library reply. Next please. So this is what a sample page from a, from a letter, a 19th century letter could look like. And so the sample page itself the digitized scan page, and an optional view of the text layer. And the last slide please. Here, we show you how we, we implemented. So this is the way we implemented highlighting for example and sleep a few in within one single document so here we I search for the personal pronoun. I, and you see here, we highlight the, the occurrences of that word, and you get some context, and our readers are actually able to, to work with this text layer, in addition to the images themselves. Okay. So now we hope that to apply this to we have applied this to 3,500 documents. And we hope to expand this in the future we have about I think 250,000 men's manuscript collections. So more to come. Thank you. Thank you. Are there any questions from the audience. Yes, please. The question was whether you curate the text in some form. Whether you do any corrections on it. Did I get that correctly. And so we do. I'm not really the person to answer this question unfortunately because I'm not part of this process. And, but we have trained a model for handwritten Norwegian 19th century, early 20th century Norwegian. And I'm not sure how then results. There might be some curation yes but I'm actually not quite sure about that. So, someone else in the audience. Excellent. I could hear it, but yeah. Okay. So any other questions or comments. If there aren't any, I think we can move on to the next presentation and Magnus, we will send you your mug, because it shouldn't be your disadvantage that you couldn't come here. And please just try to Miriam at info at recall.eu and demand your mug. Thank you. Thank you for joining us. And next up is Stefan von der Heide from content conversion specialists a German company. He will be talking about the new transcribers rest API and provide an experience reports about project with Cyrillic Ukrainian handwriting.