The Fifth Elephant 2026 Winter Edition

Submission type: Talk/Showcase proposal - 20 mins

showcase

Digitising a book is only the beginning. Across India, thousands of books, manuscripts, photographs and historical documents are being digitised, but much of this knowledge remains difficult to discover, search, cite or reuse. Researchers often have to navigate scattered repositories, incomplete metadata, scanned PDFs and language barriers. At Servants of Knowledge, we have been working on this problem through SOK Research Desk, an open-source platform that brings together digital libraries, archives and research tools. Built with Indian languages in mind, Research Desk can bring in collections from the Internet Archive, institutional repositories and local storage, make their contents searchable at the page level, and support multilingual discovery, reading, citations, annotations and collaborative OCR correction.

In this session, I will demonstrate Research Desk and share what we have learned while building an open technology stack for preserving and accessing knowledge. We will look at how collections move from digitisation to discovery, how metadata and OCR can be improved through human participation, and how open standards can help libraries and institutions work together rather than create isolated repositories. I will also discuss the practical challenges of building such systems for Indian languages, including inconsistent metadata, OCR quality, transliteration, infrastructure costs and long-term preservation. Beyond demonstrating the platform, I hope to start a conversation about how researchers, archivists, developers and institutions can work together to make our collective knowledge more accessible and useful.

Take Aways

1. From scanned documents to usable research data: Understand how open-source tools, multilingual search, metadata, OCR correction and research workflows can transform digitised collections into resources that people can actually discover, study and cite.

2. Building open knowledge infrastructure together: Explore how libraries, researchers and technology communities can create interoperable, sustainable platforms for Indian-language and cultural heritage collections without depending entirely on proprietary systems.

Audience
Researchers working with digital humanities, Indian languages, history and cultural heritage; librarians and archivists managing digital collections; open-source developers and civic technologists; data and AI practitioners working with multilingual datasets, OCR and knowledge discovery; and institutions interested in making their collections more accessible.

Bio
Omshivaprakash works at the intersection of technology, language, culture and knowledge preservation. Through initiatives such as Kannada [[Sanchaya | https://sanchaya.org]], Sanchi Foundation and Servants of Knowledge, he has been involved in building open digital resources, language technology tools and digitisation workflows that make literature, historical material and cultural knowledge accessible to the public.

His work focuses on bringing together open-source software, community participation and digital archiving to address the challenges faced by Indian languages in the digital world. He is currently involved in developing tools such as Research Desk, Repub and other open infrastructure for digitising, preserving and discovering knowledge.

Bio: https://omshivaprakash.github.io

Links

Project repository: https://github.com/servantsofknowledge/researchdesk

Research Desk: https://research.servantsofknowledge.in

Draft slides To be added. The project repository and demonstration link are available for reviewers.

One of the challenges we are exploring is how to make large, multilingual historical collections genuinely useful for research, especially when OCR is imperfect, metadata is inconsistent and resources for Indian languages are limited. We would welcome ideas and critique on improving multilingual search, creating reusable OCR training datasets, designing better research workflows and making the platform easier for smaller institutions to adopt and maintain.

We are also interested in collaborating with institutions that hold collections but lack the infrastructure or technical capacity to make them discoverable.

I would especially value critique on how to design long-term community workflows for OCR correction and manuscript transcription: how to keep contributors motivated, how to validate corrections without creating bottlenecks, and how to measure quality across multiple Indic scripts and mixed-language books. I am also interested in feedback from libraries or archives that have tried to connect local catalogues, Internet Archive records, Wikisource, Wikidata, and preservation storage into one coherent workflow. Also any support on Keeping this mission alive - through fund raising and otherwise.

I can help with: ideas, experience, technique, dataset, collaborator intro, critique, facilitation

I’d like to meet: researchers, practitioners, domain experts, tool builders, dataset owners

Topic/domain tags - digital libraries, public knowledge infrastructure, Indian languages / multilingual search, Data for good · Digital humanities · Education