Dicta Genizah Search Pro 9.11.1
Collaborative Research Platform for the Cairo Genizah
A comprehensive research environment for the Cairo Genizah, featuring a Web Platform (Dicta Genizah Search), Community Features, and full WCAG 2.0 Accessibility.
Web Access: genizahsearch.com (אתר הגניזה של דיקטה) - Search, browse, and collaborate from any browser
What's New in Version 9.11.0?
FGP transcriptions from FGP's own files
About 14,600 new Friedberg Genizah Project transcriptions and cleaner text in about 27,800 more,
now read from FGP's own files. Website and desktop.
Earlier releases: CHANGELOG.md. The README's previous What's New sections
(8.1.0 to 9.10.0) are kept in docs/archive/README_WHATS_NEW_ARCHIVE.md.
Features
Search
- Keyword, phrase and pattern search over the MiDRASH automatic transcriptions of the whole
corpus, with fuzzy and spelling-variant matching for Hebrew and Judeo-Arabic, a
Responsa mode and a tabular query builder; results span page boundaries
- Letter-level parallels: paste a passage, or several witnesses of one work, and find the
manuscripts that carry it (desktop; also via the API); Composition Search by chunks
- Lab Mode: parallel detection after Shmidman, Koppel and Porat (2016), with rare-letter
encoding for spelling tolerance
- Library filter (show only / hide), exclusion lists per tab, and export to Excel, CSV and DOCX
Reading
- IIIF viewer with high-resolution images from the NLI, Cambridge, Manchester (LUNA), JTS and
Princeton (Figgy); folio navigation with recto/verso notation; zoom and rotation; print /
save as PDF
- Oxford Bodleian support with Neubauer catalog integration, and an honest notice with a
Bodleian link when no image can be fetched
- Editions side by side: MiDRASH transcription, Princeton Geniza Project and Friedberg
editions, translations and community corrections, each credited to whoever made it, with a
Cite this page button and prominent PGP links
- Scholarly metadata: FJMS catalog records, bibliography, joins and measurements; NLI
catalog data; Hebrew translations of library metadata
Discovery
- Computed Identifications (public beta): manuscripts matched to known works, with the
evidence shown, on /computed-identifications
and in a connections panel while browsing
- Visual Genizah Atlas (preview) and a guided Start Here launchpad
Community and personal work
- Accounts, corrections, comments and discoveries (Supabase-backed)
- Personal lists, recently viewed, and cloud sync between web and desktop
- Joins Lab and the Fragment Puzzle for assembling fragments visually
- My Library (desktop): a private side-index of your own PDF, DOCX and TXT files, searched
alongside or instead of the corpus
AI and automation
- GenizahSearch in ChatGPT (an app you add from genizahsearch.com/ai;
the GenizahSearch GPT remains until 2026-12-11), a research skill for other AI assistants, and a
public HTTP/JSON API
API
GenizahSearch exposes a public HTTP/JSON API for research automation: keyword/Responsa search (POST /api/search), single-manuscript drill-down (GET /api/browse), and composition-parallels detection (POST /api/parallels). All endpoints are anonymous and rate-limited (120 req/min per endpoint per IP in the default public deployment) and return JSON in a uniform envelope.
Full reference, curl examples, and interactive Swagger UI: docs/SEARCH_API.md · genizahsearch.com/api/docs.
Getting Started
Web (Recommended)
Visit genizahsearch.com to start using Dicta Genizah Search immediately.
With an AI assistant
Add GenizahSearch to ChatGPT from genizahsearch.com/ai and ask a question, a name or a passage; no installation (the GenizahSearch GPT remains until 2026-12-11). Other assistants can use the research skill or the public API.
Desktop Installation
- Download: the installer (
GenizahSearchPro_V9.11.1_Setup.exe, about 512 MB) from the latest GitHub release
- Install: Run the installer and follow instructions
- Data Setup: The software requires the MiDRASH dataset (
Transcriptions.txt)
Antivirus Note: Some antivirus software (Avast, AVG, Windows Defender) may flag the installer as suspicious. These are false positives caused by PyInstaller packaging. See ANTIVIRUS_INFO.txt for details and solutions.
Documentation
For detailed documentation, see the docs/ directory:
Development
python -m web.main # web app on port 8081 (GENIZAH_PORT)
python genizah_app.py # desktop app
python scripts/run_local_tests.py # the test suite, in bounded lanes (never one pytest tests/ process)
Python 3.11, NiceGUI (web), PyQt6 (desktop). Read docs/architecture/OVERVIEW.md
for how the repo is organised, CLAUDE.md for project context, and
docs/guides/DEVELOPER_GUIDE.md before changing
environment variables or data paths. Release process: python scripts/bump_version.py X.Y.Z,
then the steps it prints.
Credits & Data
- Hosted & Supported by Dicta — The Israel Center for Text Analysis
- Development: Hillel Gershuni
- Data Sources:
- Stoekl Ben Ezra, D., Bambaci, L., Kiessling, B., Lapin, H., Ezer, N., Lolli, E., Rustow, M., Dershowitz, N., Kurar Barakat, B., Gogawale, S., Shmidman, A., Lavee, M., Siew, T., Raziel Kretzmer, V., Vasyutinsky Shapira, D., Olszowy-Schlanger, J., & Gila, Y. (2025). MiDRASH Automatic Transcriptions of the Cairo Geniza Fragments [Data set]. Zenodo. doi.org/10.5281/zenodo.17734473
- Princeton Geniza Project (PGP) — Curated transcriptions, translations, and metadata for Cairo Genizah documents
- Fragment of the Jewish Manuscript Studies (FJMS) — Domain classifications, scientific joins, bibliography, catalog records, joins and visual joins suggestions, and much more.
- Lab Mode Algorithm: Based on Shmidman, Koppel, and Porat (2016)
Acknowledgments:
Assisted by Claude, Gemini, and GPT.
Special thanks to Avi Shmidman, Elisha Rosenzweig, Efraim Meiri, Elazar Gershuni, Itai Kagan, Elnatan Chen, and Adiel Breuer.
