Terms of Use — Dicta Genizah Search

Effective date: 10.10.26

Who we are

Dicta Genizah Search ("GenizahSearch," "the site," "we") is a free research platform for Cairo Genizah manuscripts at genizahsearch.com. It includes a web app, a Windows desktop app, a public Search API, and AI integrations — an MCP server at https://genizahsearch.com/mcp that ChatGPT, Claude, and Gemini users can connect to, and (for as long as it lasts) the GenizahSearch custom GPT in ChatGPT.

GenizahSearch is published by Hillel Gershuni, with the support of Dicta — The Israel Center for Text Analysis.

Acceptance

By using the website, the desktop app, the Search API, or any AI tool connected to GenizahSearch (the MCP server or the GPT), you agree to these terms. If you don't agree, please don't use the service.

1. AI answers can be wrong

ChatGPT, Claude, and Gemini (through the MCP server) and the GPT can search the corpus and describe what they find, but they can also misread, mistranslate, or misattribute a manuscript, or simply get something wrong. These tools are a research aid, not a source of truth. Before relying on or publishing a claim an AI tool gave you, check it against the actual manuscript images, the transcriptions, and the catalogue records on the site.

2. Sources keep their rights

GenizahSearch brings together work from many sources, and each one keeps its own rights:

• MiDRASH — credited for the automatic transcriptions only, not for any edition, translation, or scholarly reading drawn from another source.
• Princeton Geniza Project (PGP) — transcriptions, translations, and metadata.
• Friedberg Genizah Project (FGP) — transcriptions, editions, and translations.
• FJMS and the National Library of Israel — catalogue descriptions and bibliography.
• Holding libraries (Cambridge, JTS, Oxford, the National Library of Israel, and others) — manuscript images and library metadata.

GenizahSearch's own software is licensed under CC BY-NC-SA 4.0 (non-commercial use, with attribution, shared alike), © Hillel Gershuni / Dicta. That licence covers the GenizahSearch code only, not the sources above.

When you use material from GenizahSearch, you're bound by that source's own terms and licence, and you should cite it accordingly — the site tells you, for each page, who made what you're looking at. GenizahSearch does not grant any right it does not itself hold.

3. Fair use — no bulk scraping

GenizahSearch is free to use, and we rate-limit and cap jobs so it stays usable for everyone. Don't try to download the corpus or its images in bulk, and don't try to work around the rate limits or job limits (for example, by spreading requests across many addresses or ignoring a "try again" response). If you need full-corpus access for research, contact us instead — see below.

4. Provided free, as is

GenizahSearch is provided free of charge, as is, with no warranty of any kind. We don't guarantee the site, the desktop app, the Search API, or the AI integrations (the MCP server or the GPT) will always be available, accurate, or unchanged — any of them may change or stop without notice.

Changes to these terms

We may update these terms as the service changes. The current version is always the one posted at this page, with a revised date at the top.

Contact

Questions about these terms: email [email protected], or use the support page at genizahsearch.com/support.