docRefinery vs AWS Textract

A building block or a finished workflow

Textract is a strong reading service. It detects printed and handwritten text, keeps tables intact, returns form key-value pairs and answers natural-language queries such as "What is the customer name?". It has specialised APIs for invoices and receipts (AnalyzeExpense), identity documents (AnalyzeID) and loan packages (Analyze Lending), and returns a confidence score for everything it identifies.

What it returns is the raw material. Deciding which document type arrived, which fields matter, which values to trust and who checks the rest is work your team builds on top, usually with other AWS services. docRefinery ships that work as the product: documents in, checked, structured data out.

Where Textract is the better choice

• You already run on AWS. Identity, billing, logging and data stay inside an account your security team already approves. For many organisations that outweighs every feature on this page.
• Compliance. AWS lists Textract as in scope for its SOC reports and as a HIPAA eligible service, and AWS publishes a long list of further programmes you can check service by service. docRefinery's SOC 2 Type II audit is in progress with no report yet.
• Price for plain OCR. Detect Document Text costs $0.0015 a page for the first million pages in US West (Oregon), and $0.0006 after that. If you only need text, it is hard to beat.
• Scale and maturity. It is a long-established AWS service with synchronous and asynchronous APIs for very large volumes.
• Specialised APIs. Analyze Lending classifies and splits mortgage packages, and AnalyzeID reads US passports and driving licences without templates.

What docRefinery does that Textract does not, or does differently

docRefinery fits when you would otherwise build the layer above OCR yourself. Most of these are things a Textract project writes itself, or goes without.

• You define the output. Each field has a plain-language instruction, a "Return as" name you choose as its output key, a type and a required flag, so nobody maps Textract's response onto your schema.
• One output format for every document. Every field returns text, alias, field_confidence and location, and every table returns rows, so one parser replaces separate handling for each specialised API.
• Classification and segmentation included. Long, multi-page files are handled at scale, separated into the documents inside them and classified into the right type, for every document type rather than mortgage packages alone.
• Layout memory. docRefinery learns each layout once and recalls it, which matters for monthly invoices and receipts from the same senders. Our guide to invoice data extraction walks through the fields involved.
• A review queue in the product. AWS's documentation names Textract as a built-in task type for Amazon Augmented AI (A2I), but now says A2I is no longer open to new customers and will get no new features, so a new team must build or buy that step. In docRefinery, values below a per-field threshold wait for a person, and every correction stays on the run.
• Null, never a guess. A field missing from the page comes back null, and business rules check the values before they ship.
• Governed fields. Owner, admin and member roles decide what may be extracted, from 20 document types and 785 field definitions.
• Learns from corrections. A reviewer's correction becomes a suggested fix, tested on that layout's own documents before anyone applies it.
• Runs outside AWS. Managed cloud, your private VPC, or fully air-gapped on your own servers on Enterprise.
• More languages. docRefinery supports 70+ languages and scripts; AWS lists printed text in more than 40 languages and handwriting in English.

What Textract costs for richer analysis

Prices rise with the features you turn on. In US West (Oregon), Tables cost $0.015 a page and Forms $0.05 for the first million pages, and Forms, Tables and Queries together come to $0.07. AnalyzeExpense is $0.01 a page and Analyze Lending $0.07. New customers get a three-month free tier. docRefinery is free up to 100 pages a month, and paid plans are custom, quoted to your volume.