Document intelligence
Turn Malaysian documents into structured data
Our in-house OCR and parsing reads financial statements, tables, charts and forms, and returns clean, structured data through one API.
Bank statement
Chart
Form
Structured data
JSON via APISample data for illustration
99.99%
Accuracy on data extraction
In-house
OCR and parsing, built by our team
Scans and PDFs
Original e-statements, scans and photos
100% API
Every capability available via API
Supported documents
Built for the documents Malaysian lenders receive
Statements from every major Malaysian provider, plus the supporting documents that come with an application.
Bank statements
Personal and business accounts from Malaysian banks
Credit card statements
Balances, due dates and every transaction
eWallet statements
Top-ups, payments and transfers
EPF statements
Contributions and balances
Payslips
Gross pay, deductions and net salary
Vehicle grants
Registration, owner and vehicle details
Business licences
Company and licence details
And more
New document types added on request
What it does
From upload to structured data
Every step runs automatically, so your team never retypes a document again.
OCR built for local documents
Reads original PDFs, scans and phone photos, tuned for Malaysian formats and layouts.
Document classification
Identifies each document type automatically, even when an application arrives as one mixed upload.
Data extraction
Pulls out fields and tables, from vehicle owner details to every line of a statement.
Missing document check
Flags incomplete applications straight away, before an officer starts reviewing.
Cross-document checks
Compares names, addresses and account numbers across every document in an application.
Why in-house matters
We built our own document technology, so we control the result
Accuracy on local formats
Trained and tuned on Malaysian documents, not generic global templates.
Fast to extend
New banks and document types can be added quickly when you need them.
Deeper fraud checks
Owning the technology end to end powers our proprietary DNA-level fraud checks.
Your data belongs to you
Customer data stays yours. On a private environment, you can also set your own data policy.
The engine behind Finory
Powering every Finory product
Bank statement analyzer
Parsed statements, payslips and EPF become income, affordability and cash flow insights.
Explore →Document fraud detection
Reading each file’s structure lets us catch edits a human would miss.
Explore →Transaction enrichment
Every parsed transaction is categorized, with merchants and entities identified.
Explore →Document intelligence: frequently asked questions
What is document intelligence?
Document intelligence is technology that reads documents and turns them into structured data. Finory’s in-house OCR and parsing classifies each document, extracts its fields and tables and returns the result as JSON through an API.
Which documents can Finory’s OCR read?
Finory reads bank statements from all Malaysian banks, credit card statements, eWallet statements, EPF statements, payslips, vehicle grants and business licences, with more document types added on request.
Can Finory read scanned documents and photos?
Yes. Finory’s OCR reads scans and phone photos as well as original PDFs. For document fraud detection, original e-statements give the strongest result.
Does Finory build its own OCR?
Yes. Finory’s OCR and document parsing are built in-house and tuned for Malaysian document formats. Owning the technology end to end also powers Finory’s proprietary DNA-level fraud checks.
Can Finory convert Malaysian bank statement PDFs into structured data?
Yes. Finory turns PDF statements from all Malaysian banks into structured data, including every transaction’s date, description, amount and balance, with 99.99% accuracy on data extraction. The data is delivered as JSON through Finory’s API, ready for spreadsheets, loan systems or credit models.
Why use Finory instead of a general LLM or a cloud OCR service?
It comes down to three things: consistency, cost and control. General LLMs can give different results for the same document from one run to the next, and any business that depends on document data needs the same answer every time. Cloud OCR services usually charge per page, and that adds up quickly at high volumes. If either tool misreads a document or can’t handle a new format, you have to wait for the provider to fix it, if they fix it at all. Finory builds its OCR and parsing in-house, so we fix issues and add new document types for you directly.
See your documents turned into data
Send us a few sample documents, and we’ll show you what Finory extracts.
Book a demo