Out now: Transkribus lite. Transkribus can now also be used in your browser. Transkribus lite is the web version of Transkribus with enhanced usability. Check it out here: 🙂
📣Today we hit another Milestone: 10.000 (❗️) AI-Models have now been trained with Transkribus. We are truly amazed by you, our hard working community.👏 For this, we are going to
#giveaway
10 000 credits to one lucky winner.🤞 Just retweet and like and you are in.
The new
#Portuguese
model is now public! Thanks to the
@traprinq
project, historical documents written in Portuguese can now be automatically transcribed using
#Transkribus
.
Click on the link to take a closer look 🤓
75.000 users have already signed up to work with
#Transkribus
.
Therefore, we
#giveaway
7500 credits. Just like and retweet and you're in. 🤞
We are thrilled that so many of you from all over the world have joined us to discover the past! 🤓🪶
50000
#giveaway
- Fifty thousand, yes 50000 people already use Transkribus. Today we hit this milestone.🥳Therefore we give away 5000 free Transkribus credits. Just retweet and follow us and you're in 😊
#Transkribus
#Community
The first
#OttomanTurkish
model is here! Thanks to
@SuphanKrmzltn
and the Digital Ottoman Corpora team, the model trained on Ottoman Turkish printed material from the late 19th and early 20th centuries is now public. Try it out here:
Great news: Transkribus is among the 10 finalists for the EU Horizon Impact award. We were selected from 225 applications. 5 out of the 10 will receive the award in September. Thanks to all who made this possible!
We have just reached 15 000
#Transkribus
handwritten text recognition models. This means 15 000 times someone has meticulously prepared and transcribed pages of handwritten (or printed material) as training data and hit the train button in Transkribus. This is just 🤯
Transkribus is back! 🥳 We really appreciate your great patience and all the support that we have received in the last few days! All Transkribus Services are back online and you can access your documents again.
The
#ScanTent
allows visitors at the Bibliothèque BnF
@laBnF
to take high-quality scans of
#historical
documents using just their phone. Find out how it works:
The Wolpertinger has been part of the Transkribus universe for quite some time. So we have decided incorporate our beloved Wolpi into our logo icon — what do you think?
#transkribus
#wolpi
Transkribus is awarded for the Austrian State Award for Digitization 2020: a perfect occasion to celebrate with chocolate from our friends
#alleakten
@stadsarchief
#Transkribus
now has a new Latin model! Based on Carolingian Minuscule manuscripts, the model covers a wide range of genres from biblical texts to mathematical treatises. Have a look here 👇
New Public Model available: German Handwriting M1
This is the new German Jack of all trades in Transkribus town. 🪶
It can handle a variety of German Handwritings from the 17th to the 20th century. Give it a try!
#Fraktur
#Kurrent
and
#S
ütterlin are three scripts that every
#German
historian needs to know. In this blog post, we look at the best 3
#AI
models for transcribing these 3 important scripts:
New tool online 🪶💻
Transcribe old handwriting on the go. Just drag an drop you images and let Transkribus work for you. For a start, the tool works with German Kurrent and Sütterlin. Give it a try and let us know what you think:
Transkribus is part of
#Wikisource
! To facilitate the transcription of handwritten manuscripts, Transkribus has been integrated into
@Wikimedia
's text recognition engine.
Lean more here 👇
NOSCEMUS 4.0 is now live. The public model can transcribe 15th to 19th century printed Latin and Greek text with an astonishing CER of 0.79%. Or in other words: the model is able to transcribe your printed Latin texts with more than 99% accuracy:
New
#Spanish
model available!
Coordinated by
@AlvaroCuellarDH
, this general model can
#transcribe
various materials, from medieval manuscripts to 20th-century documents.
Give it a try here:
Update for anyone working with
#historical
documents in
#Dutch
: The Amsterdam City Archives
@stadsarchief
recently released a new public "super" model for 17th-century notarial documents, based on a whopping 11,783 scans. Try it out 👇
What a lovely bunch of people 🥰 So many different backgrounds, nations and perspectives coming together here in Innsbruck is just great 🤩 And another 250 people joing remotely 🤯
The power is in your hands! Transkribus users can now learn how to train their own models for Handwritten Text Recognition thanks to our new How to Guide
Handwriting of French philosopher Michel
#Foucault
can now be transcribed automatically with around 90% accuracy - thanks to collaboration between
@Transkribus
and Foucault fiches de lecture project 📝👍
#dh
Three new, large public models are available 📜
🇩🇪 Acta 17 (extended) - For German and Latin
🇫🇷 Transkribus French Model 1
🇵🇱 Transkribus Polish Model 1
Give them a try and let us know what you think.
#Transkribus
is used by
#archives
around the world to
#transcribe
their materials and create digital versions of them. In this blog post, we take a look at how to best digitise archival materials with Transkribus, including some
#bestpractice
examples 👇
Transkribus recognises handwriting of an early modern regent - with an impressive Character Error Rate of less than 9%! 😎 Find out more and read a detailed report on the process (in German) by
@prellmartin
@UniJena
The new Transkribus web app is coming on August 30th! Check out our new article to find out what's changing and see the next steps in our journey to improve the transcription of historical documents. 👇
👑Meet "The Dutchess I." - the new
#Transkribus
AI model for Dutch! 🇳🇱 A whopping 11.7 million words of ground truth from the 17th to the 19th century, fusing models from
@NLNatArchief
&
@stadsarchief
, are the basis of "The Dutchess I." 📜
We will host a free Transkribus beginners' webinar on June 17 at 6pm CET. Tune in on YouTube to get started with Transkribus. We will talk about the most important features of Transkribus (manual transcription, layout analysis, HTR etc.)
The new Transcription Editor is now available in Transkribus Lite. There are many new and improved features such as shortcut support, view switching, annotation, and a saving reminder so you never lose your work. Try it out and let us know what you think.
We are pleased to announce that the Transkribus User Conference will be returning to Innsbruck in 2022. The conference will take place on September 29-30. 🪶
Really looking forward to seeing you all in real life (but there will be online options, too). 🤗
Boost your paleography skills with
@Transkribus
LEARN! Our e-learning site lets you practice transcribing all kinds of historical documents. You can even upload yr own documents and create learning exercises for students 👩🎓👨🎓
#twitterstorians
#archives
#dh
Another milestone for Transkribus: last week we hit the 40.000 registered users mark. 🤩 We are glad to have such a broad user base and therefore being involved in many inspiring projects . ⌨ 🙌
#keepontranscribing
66 speakers will share their experience at
#TUC22
. We are amazed that so many experts will share their experience at this year's conference. If you want to participate, too, grab one of the last in-person tickets or simply join us virtually:
We are probably only scratching the tip of the iceberg, but what is possible with HTR is already proven by some of the projects we have had the pleasure to support with our read&search application. Have a look at them:
Are you new to
@Transkribus
? Do you want to apply Handwritten Text Recognition technology to historical documents? Master the platform in just 10 steps with our updated video
Hopping in just after
#Easter
is our new
#GermanGiant
AI model for
#German
handwriting 🐰 Not to be confused with the rare breed of domestic rabbit, this multi-purpose model was trained on a whopping 15.6 million words from the 16th to the 21st century.
We're working to improve the automated recognition of historical documents laid out in tables. With the help of sample tabular documents from
@Transkribus
users, we're making good progress! 🤓 Read more:
@evimala
#history
#dh
The first model for Irish Gaelic print is public!
This model can read Irish Gaelic (Gaeilge) or Seanchló (common pre-mid-20th c.) as well as the standard Roman typeface used today.
Try it out using the link below 👇
#Irish
#Gaelic
#Seanchl
ó
Our multilingual introduction webinars continue with Spanish 🇪🇸 :
Introducción a Transkribus en español este Jueves 15 de diciembre, 10.30 CET.
Regístrese en este enlace:
Ever wondered what the difference is between OCR and HTR?📖🆚🪶
According to our R&D Lead Felix Dietrich "OCR for machine text has effectively been solved [while] handwritten text recognition remains at the cutting edge of science". Read the full article:
🎉🥳 We've reached 10K followers! To celebrate this, we're giving away 10K
#Transkribus
credits to one lucky follower! 💫
Here's how to enter:
Follow us (if you aren't already!)
Retweet this post
Reply and tell us how you'd use these credits.
Entries close on June 18th. Winner
New: Transkribus Learn, a tool for learning how to read old handwriting. Try it out on our public beta and start discovering the stories hidden in old documents!
Beta👉
I thought I was in trouble
#BNF
(I always seem to do something wrong) when the librarian approached me about my photo-taking. But in fact she was kindly offering this contraption to make it easier! The phone sits on top, so you don't have to keep repositioning...
The new Transkribus Lite 2.0 is here 🎉
Over the past few months we've been integrating as much of your valuable feedback as possible to make the browser version of Transkribus easy to use as well as powerful.
📢 The new
#Transkribus
document editor is now live! Features include a unified view for both layout and text recognition, more customisation options and enhanced tagging. Try it out now at or find out more on our blog.
🤓Transkribus Tips For Beginners🤓
#Transkribus
collections can be viewed as thumbnails or as a searchable list. Simply click on the "Switch view" button to switch between the two options 👍
"... it’s our duty as scholars to use and critically analyze deep learning not only to make cultural heritage accessible but help understand the technology and its pitfalls for our future benefit." Read the fascinating Transkribus Story of
@thist_be
here:
Transkribus 1.19.0. is out. Along with other new features, this desktop version now supports Apple M1 chips. For proper operation, please update to Java 18.
Improve your table recognition with our upcoming trainable Table Models!
After the Field Models, this is a big step forward for
#Transkribus
trainable layout
#AI
.
Learn more here:
Now online:
Drag&drop text-recognition tool for quickly recognising handwritten and printed text in historical documents.
Already available in 14 languages. 🪶🤓
Just finished our 3 million pages project with
@NA_Archief
documents. A big, big thank you to our colleagues at NA. This project made it possible to start READ-COOP! Let's see what the future brings.
The
@ScanTent
makes it to Africa!
@bixhiribarren
tested out the tent in Mali as part of the Endangered Archives Programme
@bl_eap
. This tool has the potential to help users around the world scan historical documents with a mobile phone!
@soas
#dh
#archives
With the new
#Transkribus
web app coming soon, take a look at where you can start automatically transcribing your documents - both in your collection and within the Document Editor itself.
100,000+ people use Transkribus. 🤯
We are super happy to reach this milestone and want to give something back by giving away 10x 10,000 credits.
Share your
#MyTranskribusStory
and win. 🤩
Due to the nomination for the Austrian State Award for Digitalisation we have been invited to the awarding ceremony in Vienna. 90 projects have applied for the prize of which 8 have been nominated. It was great to be part of the TV-broadcasted ceremony, even though not awarded 🥳
Exciting news! With the help of the
#AI
of
#Transkribus
it has now been possible for the first time in history to completely decipher the Adûnaic language 😮😎
Can you believe it? There are now over 20,000 tremendous
@Transkribus
users! A huge network working with Handwritten Text Recognition technology on historical collections
#history
#dh
#archives
We've been nominated for the Digital Humanities Award in the category of "Best DH Tool" 🥳 You can vote for us at before the closing date of March 26th.
@dhawards
#digitalhumanities
Tomorrow we will host a free Transkribus beginners' webinar in GERMAN at 5pm CET. Tune in on YouTube to get started with Transkribus. We will talk about the most important features of Transkribus (manual transcription, layout analysis, HTR etc.)
Är litet förälskad i den nya handskriftsläsande datamodell jag snickrade ihop i helgen. Bygger bland annat på 316407 korrlästa 1700-talsord och massa härliga entusiasters frivilliga handknackande i
@Transkribus
.
We are happy to introduce the updated version of Transkribus Lite (v 1.2.) with an improved user interface. In the new interface all features can be used via the newly added sidebar menu. Check it out yourself:
A new NOSCEMUS model is now available. Reads not only printed Latin documents but now also Greek texts from 16th C. onwards. And expans abbreviations automatically.
Are you working with documents written in Gothic script? 🤔 Join a new
@Transkribus
working group aiming to train strong Handwritten Text Recognition models that can automatically recognise all kinds of Gothic writing
#dh
#medievaltwitter
@thist_uzh
Yesterday READ turned into READ-COOP! A first European Cooperative Society as the basis for our future work with Transkribus.
Thanks to all who made this possible! It will be a big adventure!
Join us on 18 July for a Transkribus webinar where we will show you how to use the new user interface to analyse historical documents. 🗝
Register now to find out more about the features and capabilities of
#Transkribus
. 👇
Our first two
#Tibetan
models are now public!
The first model is made for the cursive script called Drutsa ('bru tsha) and was created by the
@tibschol
project. Have a look and try it out here:
👏Another public model out now: The Project REPUBLIC by
@HuygensING
has released a new public model for 17th and 18th century texts, which you can try out in Transkribus now:
More about the project in our success story:
Calling all
@Transkribus
users! Automated processing of tabular historical documents remains a big challenge. Can you help us improve the technology by providing images of tables for our experiments? Find out more:
🎉 The new Transkribus web app is officially launched!
Get started with the new
#Transkribus
platform by taking a look at the main improvements and new workspaces in this article.