01
AI-assisted reading
VEDA AI reading environment
Reading Vedic texts with support for morphology, lexical lookup, correction, annotation, and OCR.
Open reading environment →AI-assisted reading and digital infrastructure for Vedic philology
Read across texts, images, lexica, and scholarship.
VEDA AI is an AI-assisted reading environment for Vedic philology, built together with digital resources for working across electronic texts, printed editions, page images, dictionaries, linguistic annotation, and scholarship.
VEDA AI
Text
Image
Lexica
Scholarship
Reading
The project
A reading environment centred on the texts used in Vedic research.
Vedic research moves between electronic texts, historical editions, scans, dictionaries, grammatical annotation, and scholarship. VEDA AI adds computational assistance to that reading while keeping texts and sources visible.
It combines the AI reading environment with open resources for comparing, verifying, and reusing philological evidence.
Research
AI assistance grounded in the practice of textual reading.
02
Text & Image
Editions and documents
Linking printed editions and page images to machine-readable texts.
Related publications →03
Language & Knowledge
Philological infrastructure
Linking annotation, dictionaries, corpora, bibliography, and scholarship.
Related publications →Research direction
Text ↔ image ↔ dictionary ↔ scholarship
Printed evidence↔Machine-readable text↔Linguistic analysis↔Knowledge
VEDA AI & Vedic resources
Core tools and open research resources.
VEDA AI
Reading with AI, and AI reading on its own.
↗Veda Text Image Viewer
Texts, printed editions, and page images side by side.
↗Text Search
Search GRETIL texts by string or by meaning.
↗Coming soon
RAG
Searches literature for Vedic research, including papers and grammars, and answers questions with citations to sources.
Veda Image
IIIF page images of primary and secondary sources.
↗Veda Bib
The JAIBS bibliography and abbreviations.
↗EWAia Abbreviation Search
Search abbreviations used in EWAia.
↗Outputs
Publications and research outputs
2026
Transfer Between Accent Notations in Vedic OCR: Learning Visual Conventions from Parallel Text
OCR · Accent restoration
An End-to-End Workflow for OCR, Accent Restoration, and Translation of Vedic Sanskrit
Digital Humanities 2026: Book of Abstracts, pp. 111–112. DOI 10.5281/zenodo.21495909.
OCR · Accent restoration · Translation
Quantifying Text Reuse Across Three Kṛṣṇa Yajurveda Recensions: Using Multi-Algorithm Computational Collation
Proceedings of the 6th International Conference on Natural Language Processing for the Digital Humanities, pp. 41–49. DOI 10.18653/v1/2026.nlp4dh-1.5.
Text reuse · Collation
Aligned Parallel Corpus of the Vedic Saṁhitās for Machine Translation
Corpus · Translation
2025
Automatic Accent Restoration in Vedic Sanskrit with Neural Language Models
Accent restoration
Improving Classical Language Machine Translation using Supervised Fine-Tuning with Philological Commentary
Translation
Vedic Sanskrit OCR as a Bridge between Text and Image Platforms
Digital Humanities 2025: Book of Abstracts. DOI 10.5281/zenodo.19494802.
OCR
Towards Accent-Aware Vedic Sanskrit Optical Character Recognition Based on Transformer Models
OCR · Accent restoration
For the complete list, see researchmap opens in a new tab.
Updates
News
Service update
Service update
Project lead
Yuzuki Tsukagoshi塚越 柚季
Assistant Professor at the Center for Evolving Humanities, Graduate School of Humanities and Sociology, The University of Tokyo, and Assistant Professor (Cross-appointment) in the Department of Electrical Engineering and Computer Science, Tokyo University of Agriculture and Technology. Fields: Vedic linguistics and philology, digital humanities, and natural language processing.
Funding
Research support
Current projects
JST BOOST
JPMJBY25D3
Building Translation and Annotation Generation Models for Reading Classical Texts
opens in a new tabMachine translation and annotation generation for Vedic Sanskrit, with annotations grounded in scholarly sources, and a platform that brings in expert review.
JSPS KAKENHI
25K21518
Elucidating Relationships among Vedic Documents Using Multilingual and Multimodal Machine Learning Models
opens in a new tabA Grant-in-Aid for Early-Career Scientists project (Apr 2025–Mar 2030) connecting electronic texts with printed images, and primary with secondary Vedic literature, through OCR and multilingual methods.
JSPS KAKENHI
26H01258
Methodological Development for Entity Analysis in Historical Materials and Inference of Inter-document Relationships Using Language Models
opens in a new tabA Grant-in-Aid for Transformative Research Areas (Research in a proposed research area) project (Apr 2026–Mar 2028) developing methods to extract and analyse entities specific to historical materials and to infer relationships among documents.