References

A selection of our client cases

Louvre Museum

France ✧ 2025

150 000 pages

Processing of auction catalogs in the archives of the Department of Islamic Arts for research and reference.

Library of Congress

United States ✧ 2024

Handwritten information extraction in copyright forms from 1951 to 1977.

Training of a customized extraction model for 15 types of information on a representative sample of 500 forms for 10 million pages. Integration of the model into a prototype AI-assisted application dedicated to extracting this information.

International Committee of the Red Cross

Switzerland ✧ 2023

700,000 pages

The ICRC archives responds to numerous requests for information from the families of victims or prisoners of war, and has started an ambitious digitization campaign.

TEKLIA offered collaborative transcription and handwriting recognition applied to lists of French World War II prisoners.

Shoah Memorial

France ✧ 2023

35,000 index cards

Extraction of personal information from the "Fichier Juif", a collection of index cards compiled by the French police in the early 1940s.

This initiative serves as a proof of concept, using an innovative deep learning model designed for automatic key-value information extraction.

French Institute of Oriental Archaeology (IFAO)

Egypt ✧ 2022

891 pages

The IFAO holds the records of the archeological campaign conducted from 1922 to 1952 by Bernard Bruyère in Deir el-Medina (Egypt).

TEKLIA transcribed, classified, and indexed four excavation notebooks containing handwritten notes and drawings by the famous archaeologist.

Sainte-Geneviève Library

France ✧ 2022

550,000 index cards

Automatic extraction of information from index files to complete the library's digital catalogue. 29,497 call numbers were automatically extracted and identified.

5 869 pages

Processing of the Catalogue abrégé de la Bibliothèque Sainte-Geneviève, that lists most of the works held by the library at the end of the 19th century.

National Archives

France ✧ 2021-2023

800,000 index cards

200,000 pages of registers

Développement de l'application web SIMARA pour le traitement automatique de documents manuscrits.

SIMARA is used in production at the National Archive for the conversion of handwritten finding aids (registers or index cards), assisted by AI handwriting recognition and information extraction.

University of Québec in Chicoutimi

Canada ✧ 2019-2022

2.8 million pages

Développement d'une plateforme de traitement de documents et d'une méthodologie complète pour l'extraction d'informations généalogiques à partir de registres paroissiaux (1850-1920) pour le projet BALSAC.

Research projects

HikarIA

Soutenue par le ministère français de la Culture dans le cadre du plan France 2030, l'étude pilote de TEKLIA avec le Musée Guimet porte sur une collection de 280 albums photographiques japonais datant du milieu du XIXe siècle au début du XXe siècle.

This initiative aims to leverage AI-based photographic analysis tools for enhanced accessibility of these historic images, providing a rich cultural context and aiding in the promotion of digital heritage conservation.

Try HikarIA

SocFace

Based on nominal census lists digitized by archival services all over France, covering the period between 1836 and 1936, the Socface project aims to extract individual demographic and economic data on the entire French territory.

Socface est un projet de recherche collaboratif mené par l'Institut national d'études démographiques (INED), l'École d'économie de Paris, le Service interministériel des archives françaises (SIAF) et TEKLIA.

Try SocFace