How to use it
- Choose a PDF, DOCX or TXT. Text is extracted in your browser and you can see it.
- Pick a summary style, or type a question.
- Press Summarise or Ask.
What is sent, and what is not
The file itself never leaves your device. Docrift extracts the text in the browser and shows you exactly what would be sent. Only when you press Summarise or Ask does that text (up to 60,000 characters, about 20 pages) go to a language model via Vercel’s AI Gateway. The answer streams back and nothing is stored.
The model is told to use only your document and to say when the answer is not in it. It can still misread things, so check any figure, date or obligation that matters against the original.
Picking a style
Key points suits reports and articles. Section by section keeps the document’s own structure. Deadlines and action items is for letters, leases and terms, where you need to know what you have to do and by when. Plain-English explanation turns dense wording into everyday language.
Questions
Is the PDF summarizer free?
Yes, within a fair-use limit on requests per few minutes.
Does it work on scanned PDFs?
Scans have no text layer. Run OCR PDF first, then paste the text into a .txt file and summarise that.
How long a document can it summarise?
The first 60,000 characters, roughly 20 pages of dense text. Split longer PDFs and summarise the parts.
Is my document used to train AI?
Docrift does not store or log your text. It passes through Vercel AI Gateway to the model provider for that one request, under their API terms rather than a consumer chatbot’s. If a document must not reach any third party, do not use the AI tools.
Related tools
- PDF to TextJust the words, in reading order, ready to copy, search or paste somewhere else.
- Compare documentsTwo versions of a contract, and you need to know exactly what changed.
- OCR PDFFor scans and photos, where there is no text to copy.
- Split PDFPull out the pages you need, or cut a long PDF into several smaller ones.