Chat with a research paper you have to read closely
Ask what the methods actually did, where a number is reported, or what a term means in this paper's usage. Answers come from the paper in front of you and point at the page.
Questions to put to a paper
Nobody reads a paper straight through. You drop into the methods to see what was actually measured, go back for the sentence defining a term the authors then use forty times, and try to work out whether the discussion is really describing the results you have just read. Having the paper open beside a chat changes the order of that work, because you can ask while still looking at the page instead of after losing your place. Select a passage — a model specification, an inclusion criterion, a footnote about who was excluded — and ask about those sentences in particular, which spares you paraphrasing a dense clause badly in order to ask about it. Answers are drawn from the paper you uploaded, which is the difference between this and asking a general assistant what it remembers about a study.
Treat it as a way of locating things quickly rather than as a second reviewer. Ask where the sample size is stated, whether the results section supports the claim the abstract makes, or which limitation the authors concede near the end — then open the page and read it yourself before any of it reaches your notes. Page references make that cheap: the claim and its source are a click apart. There is a firm boundary at the edge of the file, too. A funding statement that is not there, a subgroup analysis nobody ran, a recruitment window the authors left vague: the useful result is learning the paper is silent on it. Each chat is tied to a single paper, so an absence in this one can never be filled in from another.
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
Research papers FAQ
Can I cite what the chat tells me?
Cite the paper, not the chat. Each answer names the page it came from, so use it as a pointer: open that page, read the sentence in context, and quote from the paper itself. Paraphrase drift, an ambiguous clause, and — on a scanned article — a misread word are all reasons a supervisor's question should meet the source rather than a summary of it.
How do I work through a reading list of several papers?
One at a time. A chat covers a single paper, so the practical method is to open each in turn and put the same short list of questions to it — design, sample, outcome measure, effect size, limitations — keeping the answers and their page numbers in your own notes. The comparison across papers stays yours to make; nothing here will tell you what a paper it has not seen concluded.
Will it explain a methods section I cannot follow?
It can explain what the paper says it did: which test was run, how the sample was split, how a variable is defined, what a term means in this paper's usage, with the page where each of those is stated. It will not rerun the analysis, check the arithmetic, or judge whether the design was appropriate. Where the methods are vague, the answer will say the paper does not specify it, which is itself worth noting.
What is worth asking first?
Ask what the paper claims to have found, then ask where the results section actually states it. Those two answers, read on the page, tell you quickly whether the abstract and the data are telling the same story. After that the useful questions are the sceptical ones: who was included, who was left out, what exactly was measured, over how long, and what the authors themselves concede at the end.
Open the paper you are meant to have read
Upload the PDF, start with the methods, and read each cited page before anything goes into your notes.