Extraction guides

PDF to JSON, by document type

One page per shape, and every number on them was produced by running the live engine at build time — not written from memory. Each links the sample PDF it was measured against.

Document	Fields	Tables	Time
Extract a commercial invoice	6	1 (9 rows)	211 ms
Extract a bank statement	4	1 (38 rows)	35 ms
Extract a packing slip	5	1 (12 rows)	12 ms
Extract a purchase order	5	1 (8 rows)	13 ms
Extract a remittance advice	4	1 (14 rows)	15 ms
Extract a utility bill	4	2 (9 rows)	11 ms
Extract a payslip	5	2 (6 rows)	10 ms
Extract a bill of lading	7	1 (6 rows)	9 ms
Extract a insurance claim form	6	1 (7 rows)	14 ms
Extract a rent roll	4	1 (24 rows)	13 ms
Extract a lab report	5	1 (7 rows)	12 ms
Extract a timesheet	4	1 (7 rows)	9 ms
Extract a credit note	5	1 (6 rows)	8 ms
Extract a expense report	5	1 (11 rows)	9 ms
Extract a quotation	5	1 (7 rows)	9 ms
Extract a balance sheet	4	1 (11 rows)	6 ms
Extract a profit and loss statement	3	1 (9 rows)	5 ms
Extract a bill of materials	4	1 (14 rows)	9 ms
Extract a certificate of analysis	5	1 (6 rows)	8 ms
Extract a work order	6	2 (7 rows)	10 ms
Extract a aged receivables report	3	1 (18 rows)	7 ms
Extract a dividend statement	4	1 (12 rows)	16 ms
Extract a price list	4	1 (16 rows)	9 ms
Extract a explanation of benefits	5	1 (5 rows)	9 ms
Extract a delivery note	5	1 (9 rows)	6 ms
Extract a inspection report	5	1 (8 rows)	7 ms
Extract a shipping manifest	6	1 (30 rows)	10 ms

Sample documents are synthetic — real ones of these types cannot be published — but the extraction output is the engine’s own, regenerated on every build.