- 01
Drop a filled or blank form.
- 02
If it may be a scan, turn on AI OCR if a scan.
- 03
Run Extract Form Fields. Text is read locally; only when it is empty does the optional OCR read the first two pages before classification.
- 04
Scan the table, download JSON, or ask a follow-up (“what is the EIN?”).
Structured data without a template
Traditional form extraction wants a template per vendor. PDF Boost AI reads the text and returns fields with a confidence tag. High means the value was explicit; low means you should look at the page.
For an image-only form, turn on AI OCR if a scan. PDF Boost AI sends downscaled images of at most the first two pages only when local extraction finds no text; it never sends the PDF itself.
Use the sample invoice to see the path: bill-to, amount due, dates, EIN. Then try your own packet.
Questions
- Can it auto-fill another PDF?
- This version exports JSON. You can paste values into the next form yourself. PDF form filling is planned as a Pro feature.
- Can it read a scanned form?
- Yes, when you turn on AI OCR if a scan. If local extraction is empty, PDF Boost AI OCRs the first two pages before extracting fields. OCR and form extraction count as separate AI runs.
How-to
How to extract fields and values from a PDF formTurn a filled application or intake packet into structured JSON without retyping every box.