Drag them into the chat. Anchor reads each file once, pictures and tables included, then answers from it every time, pointing to the exact place it found the answer.
Each answer points back to the place it came from, in the form that file has.
Lecture notes, readings, past papers. Text, tables and equations, and the graphs and screenshots inside.
Each slide's titles and points, in order.
Course outlines, handouts and notes, with their headings kept.
Plain text and Markdown notes (.txt, .md).
Flowcharts, diagrams and screenshots (.png, .jpg, .webp), turned into a written description, arrows included.
Every cell with its row and column labels (.xlsx), so number questions are answered from the cell itself.
Tap the camera button in the chat and photograph each page. A scanned PDF from your phone's own scanner app works the same way.
Take a photo per page, or choose photos you already have. Retake, reorder or remove pages before you save.
Each photo is cropped to the page and straightened, then its text is read by OCR (text recognition). A blurry or empty photo is flagged so you can retake it.
See what was read beside each photo and correct any misread word. The pages are saved as one PDF, and any answer that rests on a scanned page shows that page.
The text, tables and equations come out with their page, slide or line kept, so answers can point back to them.
Each graph, flowchart or screenshot gets a written description. Arrows are read three times and kept only where the readings agree.
Your file is split into passages and indexed two ways: by its words and by its meaning, so both exact terms and plain questions find it.
A few minutes for a typical PDF. You'll see “Added to your Library” in the chat.