Import a PDF
Upload a PDF when you want a paper, report, or document to become source material in an Atlas project. After processing, the PDF can be searched, cited, summarized, mapped, and used in grounded chat.
What this page covers
Exact upload entry point, validation, processing checkpoints, completion, and safe retry.
Before uploading
Choose a file that Atlas can extract cleanly:
- the file is a PDF
- the PDF is not password-protected
- the text can be selected in a normal PDF reader
- the file size is within your plan limits
- the document belongs in the project you currently have open.
If this is your first import, use a clean research paper or report. Save scanned documents and unusually large files for later.
Upload the PDF
- Open the project where the PDF should live.
- Open the add-source flow.
- Choose the PDF upload option.
- Select the file from your computer.
- Wait for upload and processing to finish.
Upload and processing are separate. The file may appear before it is fully ready for chat, maps, or citations.
Confirm the PDF is usable
Open the PDF from the project source list and check:
- the title is recognizable
- pages render in the viewer
- body text is present
- search can find a phrase from the paper
- a simple question about the source returns a relevant answer.
For example, ask: What is the main claim of this PDF?
If something fails
| Symptom | Likely cause | Next step |
|---|---|---|
| Upload is blocked | File type, file size, quota, or account state | Check Plans and limits |
| Processing fails | Protected, malformed, or difficult-to-extract PDF | Try a cleaner text-based copy |
| Text is missing | Scanned or image-only PDF | Use a version with selectable text if available |
| Citation page looks strange | PDF page labels differ from extracted page order | Verify the nearby text and page number together |
Next
- [Read PDFs](/docs/read-and-move through-pdfs)
- [PDF viewer](/docs/read-and-move through-pdfs)
- Generate a knowledge map