Article API
An article is more than a block of text. It has a title, an audience, a source trail, a publication state, and a destination.
Explore this guide ↗📄 Foundations · Topic field guide
PDF Article API
A PDF article workflow converts a document’s meaning into a readable draft while preserving its evidence trail. Reading order, multi-column layouts, footnotes, and tables deserve particular attention: text that looks correct in a page preview can become misleading after extraction.
Start with the right boundary
Begin with a document inventory. Identify whether the file contains selectable text, scanned pages, complex tables, or image-only diagrams. Use the least lossy extraction method available, then review the actual source page around every important claim.
A practical workflow
01. Map the document
Record page labels, section headings, and the location of tables and figures. Keep physical page numbers distinct from numbers printed inside the document.
02. Check extraction quality
Verify reading order and inspect missing characters, merged columns, repeated headers, and broken table rows before drafting.
03. Draft with page-level evidence
Attach each factual note to a page or section. Preserve qualifications, units, and exceptions when summarizing the source.
Before the next handoff
- Reading order checked on representative pages.
- Tables reviewed with their headers and footnotes.
- A draft source map that a reviewer can follow.
KEEP THIS IN MIND
Do not use OCR as a substitute for checking layout. Recognition can produce plausible words while losing the relationships between them.
Is a PDF summary the same as a faithful extraction?
No. Extraction recovers content; summarization selects and interprets it. Both stages need their own quality checks.
Take the next step
Continue with turn a pdf into an article without losing the evidence for a complete explanation, practical decisions, and a source-backed reference. To compare the neighboring parts of the process, explore the full article workflow library.