The paper is two columns.
Your reader disagrees.
A conference PDF is typeset for print: two columns, figures that float, and a text layer that copies out in the wrong order. Chaptera detects the columns, reflows the paper into a single readable thread, and keeps the figures and tables addressable so you can ask about them.
When the book does not answer your question, it tells you so instead of inventing an answer.
Drop in the PDF
An arXiv preprint, a conference paper, a scanned chapter from a proceedings volume. Column detection and OCR run automatically.
Read it as one thread
One column, adjustable type, a table of contents built from the paper’s own sections, and text you can select and search.
Interrogate it
Ask what a section claims, what a table reports, or what a figure shows. Every answer cites the page it read so you can go and check it.
Questions people actually ask
- Does it handle two-column papers?
- Yes — detecting the columns and reading them in the right order is the main thing the conversion does. Without it the text layer of a two-column paper copies out interleaved and unreadable.
- Can I trust the equations?
- On an EPUB, yes; they come from the publisher’s own markup. On a PDF they come from OCR, so treat them as a transcription to check, not a source to quote.
- What does it do with figures and tables?
- They are extracted and indexed so you can ask about them and jump to them. It will describe what a figure depicts; it will not report numbers that only exist as pixels in a plot.
- Where do the answers come from?
- Only the paper you uploaded. Each answer cites the page it read, and when the paper does not address your question it says so rather than answering from general knowledge.
What it does not do yet
You have been sold AI reading tools before. Here is where this one is weak, so you can decide before you spend an evening on it.
- Conversion quality follows scan quality. A clean digital preprint comes out clean; a photocopied scan comes out with artefacts, and the reader shows you the original page alongside.
- Equations come from OCR on a PDF, so a fraction or a subscript can occasionally come out wrong. Verify a formula before you cite it.
- It will tell you what a figure shows, but it will not read a value off a plot unless the value is printed there.
- It runs in a browser, and it is a private beta.
Free Pro through the end of the year
The first 50 invitations include free Pro through the end of the year. Invitations go out in small batches so I can actually talk to everyone. Tell me what you are trying to read and I will send one — all I ask is that you tell me what breaks.
Request Early AccessAlready have an invite? Sign in.