TextVariant Explorer is an application for exploring and comparing text variants in structured corpora. One common use case is textual criticism in theology and manuscript studies, but the same workflow applies wherever many versions of a text need to be compared: multiple editions, parallel translations, or revision histories (for example, Wikipedia). The aim is simple: make differences easier to find, compare, highlight, and discuss, without relying on ad hoc spreadsheets and manual collation. It is designed to support both close reading (inspecting individual readings) and broader exploration (spotting patterns across larger collections).

Manual textual comparison is time-consuming and repetitive, especially once the number of witnesses grows beyond a small working set. Variation is evidence, but much of the work can turn into mechanics: locating differences, aligning readings, and keeping recurring patterns in view. TextVariant Explorer is meant to reduce that friction so attention can stay on interpretation.

Exploration in TextVariant Explorer centers on two complementary modes: searching and close comparison. Search is built to support research questions, ranging from simple keyword queries to exact phrase search, boolean combinations (`AND`/`OR`), and field-based filters. The insights views support side-by-side comparison of selected readings, show differences at a glance (additions and omissions), and use plugin-driven highlighting so recurring patterns are easier to spot and communicate. For more examples and additional features, see the Exploration of Data section on Read the Docs.

Many corpora contain variation that is best treated as a single concept: spelling variants, abbreviations, translations, or inflected forms. TextVariant Explorer supports a configurable variant search that maps many surface forms to a shared label. For example, a dataset can treat "Amerika", "America", "USA", "US", and "United States" as variants of the same concept, so a single search retrieves all of them. For a field such as `name`, a query can target the label (for example, `@name:Amerika`) instead of enumerating every spelling and abbreviation.

Alongside keyword search, TextVariant Explorer can use embedding models to support semantic search. In practice, this means the tool can retrieve passages that are similar in meaning even when the surface wording differs. Several embedding models are supported inside the application, suited to different languages and research needs. The Ancient Greek Variant SBERT model from related research is supported as well; see Ancient Greek Variant SBERTs.

TextVariant Explorer is built to be extensible. Plugins can add tagging logic, UI extensions, and additional views so the tool can be adapted to the needs of a specific corpus, edition, or research workflow. See Development.

Coming Soon.

Coming soon.

If you use TextExplorer in your research, please cite it using the information provided in the CITATION.cff file.