Document Intelligence Models

Predefined models

Predefined models are ready‑to‑use AI templates designed to recognise and extract key information from the most common business documents without requiring any training or configuration. They allow organisations to immediately benefit from AI‑powered data extraction with minimal setup effort.
OptimiDoc Cloud includes the following predefined models:

1. Invoice Model
   Automatically extracts typical invoice fields such as invoice number, supplier details, dates, totals, and line‑item‑related values.

2. Contract Model
   Identifies contractual data such as parties, dates, contract IDs, signature details, and key clauses.

3. Receipt Model
   Captures information from retail and expense receipts, including totals, tax, merchant details, and timestamps.

4. Identification Document Model
   Extracts personal details from ID cards, passports, and similar documents — such as name, date of birth, ID number, and validity dates.

5. Credit Card Model
   Designed to read essential card information securely where permitted, supporting controlled workflows that require card verification.

Supported languages and document fields

General document and forms

The General Document Model is a versatile, broad‑coverage model designed to extract structured information from any type of document—especially when the document does not fit into a predefined category like invoices, receipts, or ID documents.

General Document Model Extracts

• Extracts key‑value pairs
  It intelligently identifies fields such as “Name: John Doe”, even when layouts vary.

• Extracts tables
  Recognises table structures, rows, and columns across forms and unstructured layouts.

• Extracts selection marks
  Tick boxes, checkmarks, and similar indicators are detected and included in the output.

• Extracts full text using OCR
  The model uses the underlying Read OCR engine, which supports printed and handwritten text, including multiple languages.
  

Supported languages: Document Intelligence - General document

Custom models

Custom models are trained using your labeled datasets to extract distinct data from structured, semi-structured, and unstructured documents specific to your use cases. Such custom models need to be created by the OptimiDoc Team.