termsuch

Termsuch is an open‑source search engine designed to provide fast, accurate term definitions and contextual usage across multiple languages. The project was originally launched in 2021 by a group of computational linguists at the University of Hamburg who sought to fill a gap in existing lexical resources. Termsuch’s core architecture combines a web crawler written in Python with a vector‑based retrieval backend built on Apache Lucene. The crawler indexes thousands of terminology databases, including the TEI P5, ISO 30600, and Wiktionary. After extraction, the terms are converted into a semantic vector representation using transformer models which facilitates similarity search.

A distinguishing feature of Termsuch is its ability to surface definition snippets directly in the search results. Users can filter results by language, domain (e.g., medical, legal, engineering), and even by source credibility. The interface is lightweight and can be embedded into other web services via a RESTful API, which has led to its adoption by several academic publishers.

Termsuch has been cited in research on terminology extraction and cross‑lingual information retrieval. Its open‑source license encourages community contributions; the project hosts a GitHub repository where developers can submit enhancement requests or new linguistic datasets. As of early 2025, the system serves over a million search queries per month and is integrated into at least twenty institutional repositories worldwide.