PRODUCT / PDF PROCESSING
Extract PDF data
Use POST /v1/pdf/data to extract fields that match a JSON Schema, with page citations. Omit schema for Markdown per page. Uses the text layer; ocr or an empty layer uses vision batches for selected pages, up to 500 pages. Use /v1/pdf/ocr for a searchable PDF output. Billed at the AI rate.
Read the implementation guide{
"url": "https://example.com/invoice.pdf",
"schema": {
"type": "object",
"properties": {
"vendor": { "type": "string" },
"total": { "type": "number" }
},
"required": ["total"]
}
}Choose your input
PDF + optional JSON Schema → structured JSON. Omit schema for Markdown. Failed extracts are not billed.
Understand the cost
Data extraction uses AI token pricing and applicable execution charges. Your accepted quote freezes the model and rate. Review the returned data before using it in your application.
Explore pricingBuild the handoff
Use your schema to request structured JSON, or omit it for Markdown. The extraction guide explains response shapes, page selection, and processing limits.
Explore the guideEXPLORE THE OUTPUT / ILLUSTRATIVE SAMPLE
Documents become usable data.
Give the API a schema and extract structured fields from an incoming PDF. Review the returned data in your application.
{
"url": "https://example.com/invoice.pdf",
"schema": {
"type": "object",
"properties": {
"total": { "type": "number" },
"customer": { "type": "string" }
}
}
}Invoice
INV—2026—042September 07, 2026
120 Market Street
San Francisco, CA 94103
Due September 21, 2026
USD · Bank transfer
{
"total": 8400,
"customer": "Northstar Labs"
}Illustrative preview. No API request is sent.
A useful addition to your application.
Explore the example request and build your first integration.