Text Summarizer

Paste any article, document, meeting transcript, or long email — get a one-sentence TL;DR and 3–7 bullet-point key takeaways in seconds. Free, no signup, no popups. The summary matches the source language: paste Chinese, get a Chinese summary; paste English, get English. Up to 12,000 characters per request.

How to use the summarizer

1. Paste your text into the box on the left. Minimum 100 characters so the model has enough to work with; maximum 12,000 characters per request (about 2,000–2,500 English words).
2. Hit Summarize. Most queries return in 2-5 seconds. The first request to a given text hits OpenAI; identical re-runs are served from a 30-day edge cache and feel instant.
3. Read the result. The TL;DR is one sentence under 25 words. Below it, 3–7 bullets cover the supporting points. Each section has a Copy button to grab the summary for an email, Slack message, or note.
4. Try the examples. If you don't have your own text handy, the three example buttons load a tech article, a meeting recap, and a policy document so you can see typical output.

What is the summary good for?

• Articles and blog posts — decide if it's worth a full read before committing the time.
• Meeting transcripts — generate action-oriented recaps after a Zoom call, especially for people who couldn't attend.
• Long emails — extract the one decision you're being asked to make, plus the supporting context.
• Research papers — get the abstract-of-the-abstract for skim purposes (always verify with the original).
• Policy documents — surface the actual changes hidden in legalese.
• Customer feedback — distil patterns from a long thread of comments.

Tips for a better summary

• Paste structured text. Articles with headings, paragraphs, and lists summarize better than a wall of unbroken text.
• Trim obvious boilerplate. Cookie banners, sign-up CTAs, and navigation menus add noise — remove them before pasting if you can.
• For very long documents, summarize in chunks of ~8,000 characters and combine the resulting bullets at the end. The model loses focus past ~3,000 words.
• The summarizer matches source language. If you want a translation as well as a summary, do the summary first, then translate the result separately.

Frequently asked questions

What model powers this?

OpenAI's GPT-5.4-mini, accessed via the OpenAI API. The text goes through a Cloudflare Worker that holds the API key as a server-side secret — your browser never sees it.

Why is the first request slower than the second?

The first request goes to OpenAI (2-5 seconds typical). Subsequent identical requests are served from Cloudflare's edge cache, returning in tens of milliseconds.

Can I summarize a URL instead of pasting?

Not yet — this version requires the text itself. Copy-paste from the source page for now.

Does the summary keep formatting?

No — the model returns a one-sentence TL;DR plus plain-text bullets. Markdown, links, and tables in the source aren't preserved in the output.