Guides
Field Setup
Fields define the data points LumiParser extracts from your documents. You configure fields in the Fields & Tables tab of each parser.
Custom fields are live on Pro and above
Custom fields you add extract live on the Pro plan and above. Customizing a default field (rename, disable, AI instructions, advanced reasoning) is saved and included in schema export, but doesn't yet change what the parser extracts for that field — default fields always use the built-in extraction logic for now.
Field types
Every field has a type that tells the AI what format to extract the value in.
Plain text
Free-form text. Use for names, addresses, reference numbers, and anything that doesn't fit a specific format.
Date
Dates in any format. LumiParser normalises to ISO 8601 (YYYY-MM-DD) in the output.
Number
Numeric values without currency — quantities, counts, percentages.
Amount
Monetary values. Extracted with optional currency code.
Classification
One value from a fixed set. You define the allowed values and the AI picks the closest match.
AI extraction toggle
Each field has an AI toggle. When enabled, the AI will attempt to extract a value for that field from every document. When disabled, the field is still present in the output schema but will always returnnull — useful for fields you fill in manually downstream.
AI field instructions
You can add custom instructions to guide how the AI extracts a specific field. Instructions are written in plain English and are appended to the extraction prompt for that field. Use this when:
- A field name is ambiguous and the AI picks the wrong value
- You want to normalise a value (e.g. “Always return the net amount, not gross”)
- You need to handle document-specific edge cases
Tables (Line Items)
In addition to flat fields, parsers can extract structured tables. A table is a repeating group of fields — for example, each row in an invoice line-item table might have a description, quantity, unit price, and amount. Tables appear as arrays of objects in the JSON output and as dedicated tab in the Document Inbox.
Default vs. custom fields
Fields labelled Default come with the parser template and cannot be deleted (though they can be renamed or disabled). Custom fields you add yourself can be deleted at any time. Deleting a field does not affect documents already processed — it only affects future extractions.
Reordering fields
Drag the grip handle on the left of any field row to reorder fields. The order you set here is the order fields appear in CSV exports and in the Document Inbox table view.
Was this helpful?