ParaConc Web
No corpus loaded
No search yet
Load a corpus, then search.
Click a row to select it in both panes; Shift+click for the larger context. Use the second box to require, exclude or just highlight a pattern in the aligned segment.

Load corpus

Pre-aligned texts in two languages. Formats are detected per file: one segment per line, <seg> start/stop tags (aligned by id when present), or TMX. Files for the two languages are paired in name order. UTF-8 and UTF-16 are read automatically.

Language names and locales are taken from the file; the first two languages are used.
The locale code sets word segmentation, case folding and sort order for that language (e.g. de, zh, ja). Leave it blank for language-neutral behaviour.

Save adjusted files

A new version of each input file is written; the originals are not touched. You will be asked where to save.

Sentence rules

How segments are split into sentences, for the sentence display and the "hits within one sentence" option. Rules are per language.

custom mode only; CJK full-width marks need no following space
a period after a short capitalised word (Dr., Mr., A., USA.), a single letter, or a number does not end a sentence
never end a sentence (case-insensitive)