PDF Summarizer
Import a PDF, cut it down to the part you care about and get a summary in seconds. The file is opened in your browser and never uploaded, and you see the exact text before it is summarized.
To summarize a PDF, import it, delete what you do not need (cover pages, references) and summarize the rest in sections. Humanit reads the PDF in your browser, shows you the extracted text first and returns a short summary, key points or fuller notes.
Quick facts
- Files
- PDF, text-based or scanned. Word, ODT, RTF, TXT and photos work too.
- Is the file uploaded?
- No. The PDF is opened inside your browser; only its text goes into the editor.
- Words per summary
- 1,000 words a run in your first month, then 300. Paid plans: Starter 1,200, Pro 2,500, Ultra 6,000.
- Scanned pages
- Read with OCR on a free account: 10 pages a day, 100 on paid plans.
- File limits
- 20 MB per file, up to 30 pages or photos at once.
- Cost
- Free account: 1,000 words in the first month, then 300 a month. No card.
- Summary styles
- Short (free), Medium, Long, Bullets and Key Points (PRO).
How it works
- 1Import the PDF. Drop it on the box. Text pages are read instantly; scanned pages are read with OCR on a free account.
- 2Keep only what you need summarized. Delete the cover, contents page and reference list. A summary is only as focused as its input.
- 3Choose a length. Short is free on every account. Medium, Long, Bullets and Key Points are PRO styles.
- 4Summarize in passes if it is long. Summarize each chapter or section, then combine those summaries and summarize again.
Most PDF summaries go wrong at the input
A summarizer can only compress what it is given, and a PDF gives it a lot of noise: a cover page, a table of contents, running headers on every page, figure captions, a reference list that can be a fifth of the file. Tools that summarize the whole file blind spend part of the summary describing all of that. Humanit shows you the extracted text before anything runs. Page numbers are already gone and wrapped lines are re-joined into paragraphs, so the cleanup that is left takes a minute: delete the front matter, the references and anything else you are not trying to understand.
Long PDFs: summarize in passes
Every run has a word ceiling. A free account summarizes 1,000 words at a time in its first month and 300 after that; paid plans go to Starter 1,200, Pro 2,500, Ultra 6,000. A 40-page report is far past any single request, so do what a careful reader would do. Summarize chapter by chapter, keep each summary, then paste the summaries together and summarize those. The second pass gives you an overview that is built from every part of the document rather than from whatever happened to fit. When an import is longer than your ceiling, the editor offers to keep the first part and ends it on a full sentence.
Pick the shape before you summarize
Short gives you a tight paragraph, the gist and nothing else, and it is free on every account. Medium and Long keep more of the reasoning, which matters when you will be asked about the document rather than just told about it. Bullets breaks the content into scannable points, and Key Points pulls out the claims and findings, which is the closest thing to revision notes. The four longer styles are PRO. If you only need to decide whether a document is worth reading in full, Short is enough.
Slide decks and handouts saved as PDF
Lecture slides exported to PDF are mostly fragments: a heading, three bullets, a chart label. Summarizing a whole deck works better than summarizing slide by slide, because the summarizer can connect fragments that only make sense together. Speaker notes are not part of a normal slide export; if your lecturer's explanation lives in the notes, export the deck with notes from PowerPoint or Keynote first. Charts and images are skipped, so anything that only exists in a chart will not be in the summary.
Scanned chapters and photocopied readings
Course readers and scanned book chapters usually have no text layer. The importer spots those pages and sends them to text recognition, which needs a free account and allows 10 pages a day (100 on paid plans). Recognition is good on clean printed pages and weaker on skewed, shadowed or heavily annotated ones. Skim the recognized text before you summarize, especially names, dates and numbers, because a summary will repeat a misread figure with complete confidence.
What a summary will miss
Compression means choosing, and a summarizer cannot know which detail your professor or your manager cares about. It also cannot see figures, and it treats every section as equally important unless the text says otherwise. Use a summary to decide what to read closely, to revise something you have already read, or to brief someone quickly. Do not quote from it or cite it. If a number or a claim matters, find it in the original PDF and cite the page.
Confidential documents
The PDF never leaves your device when you import it. The text you choose to summarize is sent to produce the summary, is not used to train AI models, and is kept in your account history until you delete it. For contracts, medical records or anything under a confidentiality agreement, delete names and identifying details from the editor before you run the summary, and check whether your organisation allows outside tools at all.
Who it's for
| In your PDF | What the import does | Before you summarize it |
|---|---|---|
| Text on ordinary pages | Read in your browser. Lines that wrap at the page edge are re-joined into whole paragraphs. | This is what gets summarized. |
| Page-number lines ("12", "Page 3 of 10") | Removed automatically. | Nothing to do. |
| Running headers and footers | Kept as text, once for every page they appear on. | Delete them; they add noise to every page. |
| Footnotes | Imported where they sit on the page, usually between paragraphs. | Keep them only if they carry real content. |
| Reference list or bibliography | Imported like any other text. | Delete it. It uses your word allowance and adds nothing to a summary. |
| Images, charts and figures | Skipped. Captions are text, so they come through. | Note anything important that only appears in a chart. |
| Scanned pages with no selectable text | Read with OCR. Needs a free account; 10 scan pages a day free, 100 on paid plans. | Sign in so these pages can be read, then proofread numbers. |
| More than 30 pages | The first 30 pages are read and the import tells you how many were skipped. | Summarize in sections, then summarize the summaries. |
| Plan | Words summarized at once | Single-spaced | Double-spaced |
|---|---|---|---|
| Free account, first month | 1,000 words | about 2 pages | about 4 pages |
| Free account, after that | 300 words | under a page | about 1 page |
| Starter | 1,200 words | about 2.5 pages | about 5 pages |
| Pro | 2,500 words | about 5 pages | about 10 pages |
| Ultra | 6,000 words | about 12 pages | about 24 pages |
Frequently asked questions
Can I summarize a PDF for free?
Yes. A free account summarizes up to 1,000 words at a time in its first month and 300 after that, using the Short style. Importing the PDF is free and needs no account for text-based files.
How long a PDF can it summarize?
The importer reads up to 30 pages and 20 MB. One summary covers your plan's words per run, up to 6,000 words on Ultra, so longer documents are summarized chapter by chapter and then combined.
Can it summarize a scanned PDF or a textbook chapter?
Yes, with a free account. Scanned pages are read with OCR, 10 pages a day free and 100 on paid plans. Check the recognized text for misread numbers before summarizing.
Is my PDF uploaded?
No. The file is opened inside your browser. Only the text you choose to summarize is sent, to produce the summary, and it is not used to train AI models.
Can I get the summary as bullet points?
Yes. Bullets and Key Points are PRO styles. Short, a single tight paragraph, is free on every account.
Can I ask questions about the PDF?
No. This is a summarizer, not a chat tool. It condenses the text you give it. To find a specific answer, search the PDF for it, or summarize just the section that covers it.
Does it read charts and images?
No. Images, charts and figures are skipped, though their captions come through as text. If a finding only appears in a chart, the summary will not include it.