PDF viewer
The PDF viewer is the main reading and verification surface for imported PDF sources. It lets you inspect the original document behind summaries, maps, and citations.
Core controls
| Control or state | Purpose |
|---|---|
| Scroll | Move through rendered pages. |
| Page navigation | Jump to a specific page or page range. |
| Search | Find extracted text inside the PDF when available. |
| Citation jump | Open a cited location from chat, summary, or another surface. |
| Zoom | Make text or figures readable. |
| Side panel or split view | Compare the PDF with notes, chat, or maps. |
Page numbers and anchors
PDFs can have multiple numbering systems:
- rendered page index
- printed page label
- article page number
- citation anchor
- extracted-text location.
These can differ. When verifying a citation, read the opened page and nearby context rather than relying only on a page number.
Search and extraction
PDF search depends on extracted text. Search works best when the PDF contains selectable text. Scanned documents, images, unusual layouts, or poor OCR can reduce search and citation quality.
If a PDF looks readable to you but Atlas cannot search it well, the file may be image-based or structurally difficult to parse.
Verification workflow
- Open the answer or summary citation.
- Read the cited passage.
- Read nearby paragraphs for qualifiers or contradictions.
- Check figures, tables, and footnotes if the claim depends on them.
- Revise the answer or note if the citation does not support the claim.
When to troubleshoot
Use Troubleshoot source processing when the PDF fails to process, lacks searchable text, opens incorrectly, or produces unreliable citations.