TABLE EXTRACTION

Extract tables from PDFs and images

Turn tables into rows, headers, and page-aware data without rebuilding document layout by hand.

INPUT statement.pdf OUTPUT headers[] · rows[] · CSV
{
  "headers": ["Item", "Qty", "Amount"],
  "rows": [["Widget", 3, 120]]
}

WHAT TO EVALUATE

JSON and CSV outputs

Use the format your workflow needs

Headers and rows
A practical data contract
Multi-page tables
Trace a table back to its source
Source page context
One endpoint, consistent output

INTEGRATION

Use one request in development and production.

curl -X POST https://api.startocr.com/v1/tables \ -H "Authorization: Bearer $STARTOCR_API_KEY" \ -F file=@statement.pdf

POST /v1/tables
curl -X POST https://api.startocr.com/v1/tables \
  -H "Authorization: Bearer $STARTOCR_API_KEY" \
  -F file=@statement.pdf

COMMON QUESTION

Which output should I use?

Use table extraction for spreadsheet-style data that needs a reliable row and column representation.

USAGE BASED

Tables from $30 / 1K pages

Start with representative documents, then pay only for the pages processed.