Extract Custom Fields from Any Document

If your document isn't a standard bank statement, invoice, or receipt — for example a rent roll, a trust account statement, a lease, or a custom form — use the Convert document flow at /extract/single to define your own fields once and extract them from every similar document afterward.

Opening the flow

1. Click Convert documents in the left sidebar.
2. Choose Custom document.
3. Upload the PDF you want to extract from.

Drawing fields

The document opens in a canvas viewer. For each piece of data you want to extract:

1. Click Add field.
2. Draw a rectangle around the data on the page.
3. Pick a field type:
   • Text — a single value (name, date, total).
   • Table — a repeating block of rows. Define column names and types.
   • AI prompt — describe what you want in natural language (e.g. "the property address on page 1"). Useful when the value isn't in a fixed location.

Repeat for every field. DocuClipper shows a live preview of the extracted values on the right.

Saving a template

Once you're happy, click Save template. Give it a name (e.g. "Rent Roll — Acme") and DocuClipper will remember every field position and type.

Next time you upload a document in the same layout, pick the template and extraction runs automatically — no re-drawing.

Running on multiple documents

From the Convert page, upload a batch and pick your saved template. Every document is processed against the same field definitions and the results land in a single table you can export to Excel or CSV.

Tips

• AI prompt fields are the most forgiving when the layout varies between documents. Use them for free-text answers; use rectangles for fixed positions.
• Tables need consistent column order. If some documents omit a column, add it as an empty placeholder in the template.

Related

• Extract data from bank statements
• Converting Forms (W-2, 1099, Paystubs, Tax Forms)