Guides
Creating Parsers
A parser defines how LumiParser extracts data from your documents. Each parser is tied to a specific document type and contains the fields you want to extract.
1. Choose a document type
When you create a new parser, you first select what kind of document it will process. LumiParser supports the following document types out of the box:
- Invoice — vendor invoices, contractor bills
- Receipt — point-of-sale receipts, expense receipts
- Purchase Order — POs from suppliers or buyers
- Bank Statement — monthly bank/credit card statements
2. Name your parser
Give your parser a descriptive name so you can identify it later. For example, “Vendor Invoices — UK” or “Monthly Bank Reconciliation”. The name is only shown in your dashboard and is never exposed through the API.
3. Configure your fields
After creation, you land on the Fields & Tables tab. Here you define which data points the AI should extract. Each parser comes with sensible defaults for its document type. You can:
- Rename default fields to match your internal naming conventions
- Change a field's data type (Plain text, Date, Number, Amount, Classification)
- Toggle AI extraction on or off per field
- Add custom fields for data specific to your workflow
- Drag and drop to reorder fields in the output
4. Upload a test document
Once you have your fields configured, head to the Document Inbox tab and upload a sample document. LumiParser will process it and display the extracted data on the right. Use this to verify accuracy before going live.
5. Connect to your workflow
After validating your results, connect LumiParser to the rest of your stack. You can use the API to upload documents programmatically, configure webhooks to push results to your backend, or export data to CSV for spreadsheet-based workflows.
Use a template
Not starting from scratch? Parser Templates come pre-configured with the most common fields for each document type. They save time and give you a solid baseline.
See Parser Templates →Was this helpful?