For Researchers & DH Projects

One platform. Full control. From scan to publication.

Transkribus gives you a complete, transparent pipeline for historical documents: text recognition, entity tagging, structured data extraction, and publication. Every step controllable, every result reproducible.

50 free credits every month · No credit card required

95%+
Accuracy on trained models
300+
Public models to choose from
100+
Languages & scripts supported
250+
Universities & archives as co-owners

Used by researchers at thousands of institutions

A controllable pipeline for every research question

Each step is transparent, configurable, and built for academic rigour.

Text Recognition & AI Training

Get accurate transcriptions of handwritten and printed documents. Use a public model or train your own on your specific script, language, and hand — then retrain and improve as you go.

Train custom text recognition models on your corpus
Train layout models for complex page structures
300+ public models for common scripts and periods
Compare model versions, track accuracy improvements

For advanced workflows

Integrate Transkribus into your research infrastructure

Use the Transkribus API to run recognition, training, and extraction programmatically. Pipe results directly into your databases, analysis tools, or publication pipelines.
REST API with full model access
Batch-process thousands of pages unattended
JSON/XML output for any downstream pipeline
response.json
{
  "processId": 47725,
  "status": "FINISHED",
  "pages": 1,
  "content": {
    "text": "Dear Sir, I hereby confirm\nthe delivery of 200 units.",
    "regions": [
      {
        "id": "region_1",
        "type": "paragraph",
        "lines": [
          { "text": "Dear Sir, I hereby confirm" },
          { "text": "the delivery of 200 units." }
        ]
      }
    ]
  }
}

Built for large-scale research

Run projects at scale without thinking about IT

Transkribus handles the infrastructure so you can focus on research. Manage thousands of pages, coordinate teams across institutions, produce training data systematically, and tailor models to your specific corpora.
Manage large collections and coordinate distributed teams with role-based access
Systematically produce and curate training data across your corpus
Train, compare, and refine models tailored to your specific scripts and hands
Zero infrastructure overhead — runs in the browser, hosted and maintained for you

Trusted by researchers worldwide.

From national archives to university departments, see how researchers use Transkribus to unlock their collections.

EUAT

Built on trust, powered by community.

Transkribus is built and hosted in Europe by a cooperative. Your data is handled with care, and the platform keeps evolving thanks to the community behind it.

Your data stays yours

Full ownership. Delete anytime.

Hosted in Europe

All processing on our own servers in Austria. GDPR-compliant. No Big Tech dependencies.

Cooperative, not a startup

250+ co-owners. Built for long-term sustainability, not a VC exit. Your research infrastructure won't disappear.

Start building your research pipeline today

Join thousands of researchers who use Transkribus to turn historical documents into structured, publishable data.

50 free credits every month · No credit card required

200M+Pages processed
500K+Users worldwide
95%+Accuracy on trained models