Parse PDF API
How PDF parse requests fit into Smart PDF RAG workflows using task responses, status checks, JSON, Markdown, and source metadata.
Submit a parse request
A parse request sends the PDF into the Smart PDF RAG pipeline. The application should associate the request with its own document id or user workflow so completed output can be stored and retrieved later.
Track the task
Async processing lets the API return quickly while parsing continues. Poll or otherwise check task status until the document is complete or requires attention.
Use structured results
Completed output should be stored with source metadata. JSON supports typed elements and coordinates, while Markdown is convenient for embedding and model ingestion.
Implementation checklist
- 1Upload or submit the PDF
- 2Capture the task id
- 3Check task status
- 4Fetch JSON and Markdown results
- 5Store source page and bbox metadata
Questions about Parse PDF API
Why use a task flow?+
Real PDFs can be large, scanned, or complex. A task flow avoids long browser or server requests while parsing completes.
What makes output RAG-ready?+
RAG-ready output preserves headings, reading order, page references, tables, Markdown, and metadata for citations.