Chat with a PDF and check the answer
Upload a PDF and ask questions in the chat beside it. Answers come from your document alone, and they name the page they were taken from, so you can go and read it for yourself.
Questions to ask any PDF
Searching a PDF and chatting with one solve different problems. Ctrl+F matches the string you typed: get the wording wrong, or ask about something the document expresses in its own vocabulary, and you get nothing back, or forty hits you then have to read yourself. A chat works from meaning instead. You can ask where the second method is explained, what a section concludes, or why the author rejects an alternative, without knowing which words the document uses for any of it, and what comes back is a written answer rather than a list of positions in the file. The document stays open beside the chat while you do this, so you are reading the page and asking about it in one window, rather than moving between a reader and an assistant that has never seen your file.
The awkward part of any AI answer is knowing whether to believe it, so these answers are built to be checked. Page references arrive inside the answer as chips; press one and the viewer scrolls to that page, which turns verification into a glance rather than a re-read. A longer answer usually carries several of them, marking the places it drew on. You can work in the other direction too: select a passage in the viewer and ask about those exact sentences, so there is no ambiguity about which part of the page you mean. And the assistant works only from the file you uploaded, so a question the document does not cover gets a plain admission instead of an answer assembled from general knowledge. An answer you cannot trace to a page is one you should not be using.
What you get, on every document
- Answers that cite the page
- Replies name the pages they came from, as chips that scroll the viewer straight there so you can read the source yourself.
- Grounded in your document
- It answers from the file you uploaded and says so plainly when your question is not covered, rather than inventing something plausible.
- PDFs up to 25 MB
- One document per chat, shown beside the conversation with real text selection — highlight a passage and ask about that exact part.
- Scans read automatically
- Pages with no selectable text are read with OCR instead, page by page, so a scanned or photographed document still works.
- Six interface languages
- English, Bangla, Spanish, Indonesian, Arabic and German — and replies come back in whichever language you ask in.
- Turn it into practice
- The same document can become practice multiple-choice questions, spaced-repetition review and mock exams once you have read it.
Other ways to read with AI
Do more with your documents
Chat with PDF FAQ
How is chatting with a PDF different from searching it?
Search matches the exact string you type, so it fails when the document words an idea differently from you, and it hands back a list of hits to read rather than an answer. A chat works from what a passage means, so you can ask which part deals with a topic without knowing the heading it sits under. The page reference then puts you where a search would have taken you, if you had guessed the right word.
What happens when I click a page reference?
The viewer scrolls to that page in the document beside the chat, so you can read the sentence the answer came from in its own context. A short answer often carries one reference; a longer one usually carries several, marking the places it worked from. The point is that checking costs a click rather than a re-read, which makes it cheap to be sceptical.
What happens if the answer is not in my document?
You are told so. The assistant is limited to the file you uploaded, so when your question is about something the document never covers, it says the document does not cover it rather than filling the space from general knowledge. That is worth testing early: ask something you know is outside the file, see what comes back, and let that set how much weight you put on answers you cannot check as easily.
Can I put more than one PDF into a chat?
No. Each chat is tied to a single document, which is also what makes a page reference unambiguous: there is only one file it can be pointing into. If you have several files to get through, open each one in its own chat. Nothing from one chat is carried into the answers in another, so an answer is always about the document you are looking at.
Open a PDF and put a question to it
Upload a document, ask the first thing you want to know, then click the page reference and see whether the answer holds.