Transkribus transcription platform

Municipality of Amsterdam

This algorithm has a low impact. It enables historical handwritten documents to be searched by word.

Last change on 2nd of July 2026, at 14:58 (CET) | Publication Standard 1.0

• Publication category: Other algorithms
• Impact assessment: Field not filled in.
• Status: In use

General information

Theme

Culture and Recreation

Begin date

2022-07

Contact information

[e-mail]

Responsible use

Goal and impact

Making historical handwritten documents digitally accessible and searchable for researchers and other interested parties. No impact.

Considerations

Making historical research easier. This gives more people access to historical source material.

Human intervention

The AI models were trained within the Transkribus tool by staff at the City Archives. The machine-read texts (HTR) were not subsequently corrected by humans, which means there may be errors in the recognised characters

Risk management

The risks are low. The City Archives do not process non-public documents using HTR. Transkribus originated from an EU Horizon 2020 programme and was subsequently developed into a European cooperative with a large number of international heritage institutions as members. All data and metadata are hosted on European servers and comply with the GDPR and AVG. 

Operations

Data

Transcriptions and Ground Truth

The dataset contains machine-read transcriptions and Ground Truth (training data) from historical manuscripts held in the notarial archives, the Public Works archive and the public section of the Civil Registry. New scans are added periodically using HTR. The training data consists of tens of thousands of transcriptions, produced by volunteers and staff at the Amsterdam City Archives.

Technical design

Using machine learning and Handwritten Text Recognition (HTR) techniques, AI models are trained to recognise handwriting, including both 17th-century and more modern examples.

Model architecture
The HTR is implemented using various specific and generic AI models within Transkribus, utilising convolutional neural networks and transformer neural networks. 

External provider

Transkribus