Sentences are sorted by length, with 50 sentences per page.
Sentences with audio are shown first, followed by sentences without audio.
Only sentences from the Tatoeba Corpus that have been proofread are included.
I only included word families that had 18 or more sentences in my proofread subset of the Tatoeba Corpus.
Warning! The Tatoeba Corpus has errors. (Read details.)
I have limited my use to English sentences that I have personally proofread. However, some of the translations may be wrong or were written by non-native speakers and sound very unnatural.
The NGSL Rank is from the NGSL Stats & Frequencies to 34K file found on the A New General Service List website. I renumbered the "Sup" and NAWL so they would be sequential after NGSL.
The NGSL Category shows the 3 categories by the NGSL group, plus "off list" if on the 34K list, or "other" if not.
This project is still under construction.
Currently, there are slightly over 3,000 sets of pages with sentences focused on words and their families.
These pages are being built from a collection of over 300,000 English sentences. 123,378 of these sentences have audio.