This is a language model reading your document — it can misjudge or misphrase things.It is not a search index or a citation tool: it reads the text and writes a summary in its own words, the same way a person skimming the document would, except faster and occasionally wrong about emphasis. For anything you'll rely on — a contract clause, a figure, a decision — read the actual document, not just the summary. The AI runs on our own hardware; your file is never sent to a third-party AI API, and it is only ever handled in memory here, never written to our disk.
By ShinobiTools Team · Last updated: August 2026
A report you don't have time to read in full, a contract you want the gist of before the meeting, a long PDF someone sent you with no explanation — upload it, pick Short, Medium or Long, and get a summary back in a few seconds.
The document's text is extracted page by page and handed to a language model running on our own hardware, with an instruction to only use what's actually in the document and to write a summary of the length you picked. It is not a fixed algorithm pulling out the "most important" sentences verbatim — it reads and rewrites, the way a person summarizing a document for a colleague would. That also means two summaries of the same document, or the same document run twice, won't come out byte-for-byte identical.
A scanned page is a picture, not text — there is nothing for the AI to read until it has a text layer. This tool detects that case and points you to Image to Text instead of returning an empty or made-up summary — that tool reads a scanned PDF with OCR and gives you the words back directly, since we can't yet add a text layer to a PDF you already have. If your document is mostly a text layer already (a report, an invoice, an exported document), this works directly.
Your PDF is processed entirely in memory on our own machine and handed back in the response — it is never written to our disk for this tool, and nothing is kept afterwards. The extracted text is sent to a language model running on our own hardware, not to an outside AI provider.
Want to ask specific questions about the document instead of reading a summary of the whole thing? Chat with PDF answers questions about the same text. Need the structure — headings, lists — kept intact instead of condensed into a summary? PDF to Markdown does that. Handed a scan with no text layer at all? Image to Text pulls the words out with OCR — we can't yet add a text layer to a PDF you already have.
Upload the PDF, pick Short, Medium or Long, and press Summarize. An AI model reads the document's text and writes a summary in a few seconds to a minute.
Yes to both. This is a language model reading your document and writing a summary in its own words, not a fixed extraction of sentences. It can misjudge what matters or phrase something imprecisely. For anything you'll rely on — a contract, a decision, a number — read the actual document, not just the summary.
Only if it already has a text layer. A scanned page is just a picture, and there is no text for the AI to read — and we can't yet add a text layer to a PDF you already have. Image to Text reads a scanned PDF with OCR and gives you the words back as copyable text instead.
No. The AI model runs on our own hardware. Your file is not sent to OpenAI, Google or any other outside AI provider.
Short is 2-3 sentences, Medium is a short paragraph, and Long covers the main points and key details across a few paragraphs. All three read the same document text — only how much the AI is asked to write changes.
25 pages per file. There is no account, no watermark and no daily cap.