It counts how often each meaningful word appears in your text, ignoring filler words like "the" and "and", then scores every sentence by the words it contains. The highest-scoring sentences are kept in their original order. Sentences near the start get a small boost, since most writing states its point early.

It reuses it. This is an extractive summarizer, so every sentence in the output appears verbatim in your input. That is a deliberate choice: summarizers that generate new wording read more smoothly but can state things the source never said, which is a poor trade when you are summarizing something you need to be accurate.

No practical limit. The whole thing is straightforward text processing rather than a neural network, so even a very long report is summarized in milliseconds without loading your device's memory.

No. The summarizer is a few dozen lines of JavaScript running inside the page you already loaded. Your text never leaves your browser, is never logged, and there is no model or service on our side that could be trained on it.

Because it is fast enough not to need one. There is no model to download and no request to send, so results appear as fast as you can paste. Adjusting the Short, Medium, or Long setting re-summarizes instantly too.

It works best on structured prose — articles, reports, essays — where sentences carry self-contained points. It does less well on dialogue, poetry, bulleted fragments, or text where the meaning depends on the order of every sentence, because pulling sentences out of such writing loses the thread.