Account

What new tool do you want?

Text & Writing

Text Summarizer

Shorten long articles, essays and reports to their key sentences — pick the summary length and get bullet points or a paragraph. Works offline in your browser.

How to use

  1. 1Paste the text you want to summarize.
  2. 2Choose how long the summary should be, and whether you want bullets.

How it's calculated

Extractive summarization: each sentence is scored by how often its meaningful words appear across the whole text (common words like 'the' are ignored), with a small bonus for the opening sentence. The top sentences are kept in their original order.

Frequently asked questions

Does it rewrite the text?

No — it selects the most representative sentences verbatim, so nothing is invented or misquoted. Edit the result if you need your own wording.

Is my text uploaded to an AI service?

No. The summary is computed entirely in your browser; your text never leaves your device.

What works best?

Informative writing — news, reports, essays, documentation — of at least 8–10 sentences. Dialogue and poetry summarize poorly.

A worked example that shows how it tells signal from noise

Feed it eight sentences mixing a genuine topic (solar panel costs and adoption) with unrelated asides ("My neighbor painted his fence blue," "The weather was pleasant on Tuesday") and a Short summary correctly keeps exactly the two most topically connected sentences — about photovoltaic cells converting sunlight and cheaper panels enabling more homes to generate power — while discarding the fence and the weather entirely. Nothing about the algorithm "understands" solar panels as a topic; it works because content words like "solar" and "panels" repeat across multiple sentences in the real topic while "fence" and "weather" appear exactly once each, so sentences built from frequently-repeated words score higher.

Why this is extraction, not generation

This summarizer never writes a single new word — it's purely selecting whole sentences, verbatim, from your original text and dropping the rest. That's a deliberate and meaningfully different approach from an AI summarizer that paraphrases: nothing can be misquoted, hallucinated, or subtly altered in meaning, because every sentence that appears in the output appears character-for-character in the input. The tradeoff is that a summary can read a bit choppy since consecutive sentences in the output weren't necessarily consecutive in the source — you're trading smooth prose for a hard guarantee of faithfulness to the original wording.

Why the formula gives the opening sentence a scoring boost

Beyond raw word-frequency scoring, every sentence's score gets a flat +0.35 bonus specifically if it's the very first sentence of the text — a deliberate nod to how informative writing is actually structured: news articles lead with the most important fact, essays open with a thesis, reports start with an executive summary. Without that bonus, a well-written opening sentence that uses mostly unique, specific wording (not yet repeated anywhere else in the piece) could score artificially low on pure word-frequency alone and get cut, even though it's often the single most important sentence in the whole piece.

TL;DR never keeps more than two sentences

Summary length opens on short, about a quarter of the sentences, and never fewer than one. Medium is about 40 percent and long about 60 percent, again at least one sentence. TL;DR keeps at most two sentences, and about a tenth of the passage when that tenth is smaller. Bullet points starts off. Turning it on puts a dot in front of each kept sentence. A passage with no sentences comes back empty. I use TL;DR when I want a hard cap of two lines, and short when a quarter of a long article is still the right size.

Related tools

More: Text & Writing