Schema Builder

What is a schema?

A schema defines the fields you want DocuStack to extract from your documents. Instead of manually reading through pages, you describe the data you need — and DocuStack’s AI handles the rest.

For example, a medical claims schema might include:

Field	Type	Description
`patient_name`	Text	Full name of the patient
`date_of_service`	Date	Date the service was provided
`diagnosis_code`	Text	ICD-10 diagnosis code
`total_billed`	Currency	Total amount billed
`provider_name`	Text	Name of the healthcare provider

How the schema builder will work

The schema builder will provide a visual interface for creating and managing extraction schemas:

Create a schema — Give it a name and description, then choose the document type it applies to.
Add fields — Define each field with a name, data type, and optional description to guide the AI.
Set validation rules — Mark fields as required, add format constraints, or define allowed values.
Test against a sample — Upload a sample document and preview extraction results before deploying.
Deploy — Apply the schema to incoming documents for automated extraction.

Supported field types

DocuStack schemas will support the following field types:

Type	Example	Use case
Text	`"Jane Doe"`	Names, descriptions, free-form text
Number	`42`	Counts, identifiers, numeric values
Currency	`$1,250.00`	Financial amounts with currency symbols
Date	`2025-03-15`	Dates in various formats
Boolean	`true`	Yes/no fields, checkboxes
List	`["item1", "item2"]`	Repeated or multi-value fields

Best practices

Be specific in field descriptions — The more context you provide, the more accurate the extraction. Instead of “amount,” use “total amount billed to the insurance company.”
Start simple — Begin with a few high-value fields and add more as you validate results.
Use sample documents — Always test your schema against representative documents before production use.
Version your schemas — As document formats evolve, create new schema versions rather than modifying existing ones.