Learning with Texts vs LinguaCafe

Learning with Texts (LWT) is a language-learning web app for reading texts you bring and building vocabulary from them. It divides uploaded texts into sentences and words, then shows saved terms and their learning status in later readings. You can save words or expressions of two to nine words, along with translations, optional romanization, and an example sentence. Practice options include term tests, context-based tests, Massive-Context Cloze Deletion tests, and a statistics page. LWT can associate an MP3 audio URL with a text for listening while you read, but it has no audio-file import or upload function. Dictionary links are configurable; terms can be imported from CSV or TSV, exported as TSV, or sent to Anki for cloze tests. LWT is free, with offline access and interactive practice listed. It requires a modern browser and a local web and database server with PHP. Installation instructions cover Windows, macOS, and Linux, though the maker says its macOS instructions and app could not be tested. Texts and terms using Unicode characters above U+FFFF are unsupported.

LinguaCafe is a free, open-source web app for learning languages through reading and vocabulary review. It is self-hosted, so users can run it on their own machine or private web server and control their data. Reading material can come from plain text, text files, ebooks, YouTube subtitles, subtitle files, Jellyfin external subtitles, or webpages. Readers can look up words and phrases, save translations, mark items as known or ignored, and track vocabulary by learning level. Spaced-repetition reviews follow an approach similar to the Leitner system, and highlighted cards can be exported to Anki or other SRS software. DeepL translations are available for lookups when enabled. Text-to-speech uses browser SpeechSynthesis, with voices dependent on the browser. The project lists many supported languages and supports Progressive Web App installation on mobile devices; it works in browsers on screens at least 340 pixels wide. Setup uses Docker Compose. The project supports x64, with an additional step for Apple silicon, but other Armv8 devices such as Raspberry Pi 3 and newer do not currently work. One user per server is supported.