Which paper sources are supported?

arXiv, PubMed, DOI links (doi.org), SSRN, Nature, Science, IEEE, ACM, Springer, Wiley, bioRxiv, medRxiv, and most open-access publishers. You can also drop a PDF file directly.

How does PDF upload work?

Your PDF is processed locally in the browser — text is extracted on your device, then only the text is sent for summarization. The file itself never leaves your computer.

What are the limits for PDF upload?

Maximum 15 MB file size and up to 50 pages. For most academic papers, that covers the entire document. Larger files are summarized from the first 50 pages.

Does it handle long papers?

Papers longer than 50 pages are truncated to the most relevant sections. The summary still captures abstract, key methods, and results even from truncated input.

Yes, within the current web quota bucket (3 summaries per day shared across all sources). The Telegram bot offers additional credits.

Paper Summarizer

Summarize Research Papers and PDFs

Paste a paper link or drop a PDF file and get the key findings distilled into a structured summary. No signup required.

arXiv, PubMed, DOI. 3 free/day.28+ academic domains supported

Output format3/3 left today

Best for arXiv, PubMed, DOI, SSRN, and open-access paper links.

Try SummaryBot in Telegram|Open in Dashboard

arXiv, PubMed, DOIResearch-awarePDF upload4 output formats

Academic Papers

URL or PDF upload, structured summary out

Paste an arXiv, PubMed, or DOI link and the tool fetches the paper content automatically. Or drop a PDF directly — text is extracted locally in your browser, nothing leaves your device.

PDF processed locally. Up to 50 pages, 15 MB max.

See the difference

One link or one file, structured insight out. Works with any open-access academic paper.

arXiv Paper

arXiv · 2301.07041

Self-improvement in reasoning models via correct chain-of-thought outputs

Summary

The paper shows that reasoning models can self-improve by training on their own correct chain-of-thought outputs, reducing dependency on human-labeled data. Key result: 3-5% accuracy gains on math benchmarks without new training data.

PubMed Study

PubMed · meta-analysis

Intermittent fasting vs caloric restriction: weight loss and adherence at 12 months

Summary

The meta-analysis of 47 trials found that intermittent fasting produced comparable weight loss to caloric restriction at 12 months. Key finding: adherence rates were significantly higher in the fasting group, suggesting better long-term viability.

Uploaded PDF

PDF · 78 pages · uploaded

Dropped a local PDF on transformer architecture optimization for edge devices

Summary (first 50 pages)

The paper proposes a distillation method that reduces transformer inference cost by 60% on mobile hardware. Key contribution: a dynamic attention pruning layer that adapts at runtime. Extracted from the first 50 of 78 pages — abstract, methods, and ablation results covered.

How to use it

Paste an arXiv, PubMed, DOI, SSRN, or any academic paper URL — or drop a PDF file.

The tool fetches the paper content or extracts text from your PDF (up to 50 pages).

Get a structured summary highlighting the key findings, methods, and implications.

This tool handles

PDF files up to 15 MB, parsed in your browser via unpdf — text never leaves the device until extracted
Research papers, white papers, eBooks, business reports, technical documentation, lecture slides
Multi-section documents — summary preserves section structure when the PDF has a clear table of contents
Bullet, paragraph, key-takeaways, and TL;DR output formats with strict length caps
Output language matches the user's interface language
Document title (from the first page or PDF metadata) included in the summary header

Not in scope

Scanned image-only PDFs without an OCR layer — there is no extractable text; run OCR first (Adobe Acrobat, Tesseract)
Encrypted / password-protected PDFs — unpdf can't unlock them
PDFs over 15 MB — split the document or extract the relevant section first
PDFs where the layout matters as much as the text (visual diagrams, comic books, infographics) — the summary loses the visual

PDF summarisation works on text — not on layout, not on figures, not on equations rendered as images. The deep-dive below covers what survives the extraction, when an OCR pass is needed first, and how to scope the file for the best summary.

Frequently asked questions

Related Tools

AI Search

Search the web with AI-powered answers and source citations.

Grammar Checker

Fix grammar and spelling in any text.

Need to summarize papers on the go?

SummaryBot in Telegram handles the same links with credits and sharing.

Try SummaryBot in Telegram