API Category · Document Intelligence
Send a photo or a scan and get the fields back as JSON. Smart OCR knows an Aadhaar from a PAN card and a payslip from a bank statement, so your systems receive clean, structured data instead of a wall of text.
+ more in this category
Trusted Across Indian BFSI
Banks, NBFCs, housing finance companies and co-operative banks run verification, Aadhaar masking and document workflows on FinHub, inside their own compliance and audit requirements.
Prüfens Advisory
In short
Smart OCR is FinHub's AI-based optical character recognition API for financial services. It reads Indian identity, banking, income and business documents and returns the fields on them as structured JSON, with quality and tamper signals on supported documents. It tells you what a document says. Whether that is true is a separate question, answered by a verification call.
- Reads identity documents, bank and income documents, and business certificates
- Returns named fields, ready for your onboarding journey or loan origination system
- Checks the capture before extraction, so a poor photo is retaken while the customer is still there
- Hosted in India
31 APIs
Every API in Smart OCR
One integration, one authentication and one response shape across the whole category.
Reference
What Each API Returns
The endpoint and the fields that come back, so you can size the integration before you write any code.
| API | Endpoint | Key fields returned |
|---|---|---|
| Aadhaar Masking | POST /v1/documents/aadhaar/mask | Masked document (last 4 digits visible) · Detected-number locations · Masking audit record |
| Document OCR | POST /v1/documents/ocr | Structured fields (name, DOB, number, address…) · Document classification · Confidence scores · Tamper / authenticity signals |
| PAN OCR | POST /v1/ocr/pan | PAN number · Cardholder name · Date of birth · Document quality score · Tamper/quality flags |
| Aadhaar OCR | POST /v1/ocr/aadhaar | Aadhaar number (maskable) · Name · Date of birth · Address · Document quality flags |
| Passport OCR | POST /v1/ocr/passport | Passport number · Name and gender · Date of birth · Address · Expiry date and place of birth |
| Voter ID OCR | POST /v1/ocr/voter-id | EPIC number · Name · Date of birth · Document quality score · Tamper/quality flags |
| Driving License OCR | POST /v1/ocr/driving-license | Licence number · Name · Date of birth · Expiry date · Document quality flags |
| Bank Statement OCR | POST /v1/ocr/bank-statement | Account holder and account number · IFSC and bank name · Transaction date, description and amount · Running balance · Document quality flags |
| Bank Cheque OCR | POST /v1/ocr/bank-cheque | Account number · IFSC code and bank name · Payee name and cheque number · Document quality flags |
| Cancelled Cheque OCR | POST /v1/ocr/cancelled-cheque | Account number · IFSC code and bank name · Account holder name (if printed) · Document quality flags |
| Salary Slip OCR | POST /v1/ocr/salary-slip | Employer name and employee ID · Earnings breakdown · Deductions breakdown · Net pay and pay period · Document quality flags |
| ITR OCR | POST /v1/ocr/itr | PAN and filer name · Assessment year · Gross total income · Tax paid and refund details · Document quality flags |
| Invoice OCR | POST /v1/ocr/invoice | Vendor name and GSTIN · Invoice number and date · Line items and total amount · Tax breakup (CGST/SGST/IGST) · Document quality score |
| Balance Sheet OCR | POST /v1/ocr/balance-sheet | Total assets and liabilities · Equity and reserves · Key line-item breakdown · Reporting period and currency · Document quality score |
| Utility Bill OCR | POST /v1/ocr/utility-bill | Customer name and address · Service provider and account number · Bill date and due date · Billed amount · Document quality score |
| Insurance Policy OCR | POST /v1/ocr/insurance-policy | Insurer name and policy number · Policyholder name · Sum assured and premium amount · Policy start and expiry dates · Document quality score |
| Mutual Fund OCR | POST /v1/ocr/mutual-fund | Fund house and scheme name · Folio number and investor name · Unit balance and NAV · Current market value · Document quality score |
| Passbook OCR | POST /v1/ocr/passbook | Account holder name · Account number and IFSC · Bank and branch name · Latest balance and transaction date · Document quality score |
| Pension Certificate OCR | POST /v1/ocr/pension-certificate | Pensioner name and PPO number · Sanctioning authority · Pension amount and category · Date of sanction · Document quality score |
| Death Certificate OCR | POST /v1/ocr/death-certificate | Deceased name and registration number · Date and place of death · Issuing municipal or registrar authority · Date of registration · Document quality score |
| Shop Establishment OCR | POST /v1/ocr/shop-establishment | Business name and registration number · Registered address · Issuing authority and state · Validity or renewal date · Document quality score |
| KYC Form OCR | POST /v1/ocr/kyc-form | Customer name and address · PAN, Aadhaar or other ID numbers entered · Date of birth and contact details · Signature and form completeness status · Document quality score |
| Loan Agreement OCR | POST /v1/ocr/loan-agreement | Borrower & lender name · Loan amount & currency · Tenure & interest rate · Repayment schedule · Field-level confidence scores |
| Legal Document OCR | POST /v1/ocr/legal-document | Full extracted text · Parties & signatories · Key dates & clause references · Document type classification · Processing status (sync/async) |
| Udyam Certificate OCR | POST /v1/ocr/udyam-certificate | Udyam Registration Number (URN) · Enterprise name & type · Enterprise category (micro/small/medium) · Major activity · Registration date |
| GST Certificate OCR | POST /v1/ocr/gst-certificate | GSTIN · Legal & trade name · Business constitution · Registration date · Principal place of business |
| Certificate of Incorporation OCR | POST /v1/ocr/incorporation-certificate | Corporate Identification Number (CIN) · Registered company name · Company type & category · Date of incorporation · Registrar of Companies (RoC) jurisdiction |
| Credit Card OCR | POST /v1/ocr/credit-card | Card number · Cardholder name · Expiry date · Issuing network · Field-level confidence scores |
| Document to Text OCR | POST /v1/ocr/document-to-text | Full extracted text · Per-line/word confidence scores · Detected language · Layout metadata |
| Document Quality Check | POST /v1/ocr/document-quality-check | Overall quality score · Blur, glare & lighting flags · Crop / edge-detection status · Resolution assessment · Retake recommendation |
| Document Identification | POST /v1/ocr/document-identification | Detected document type · Classification confidence score · Side detection (front/back, where applicable) · Unsupported/ambiguous flag |
What Smart OCR Reads
Identity, income, banking and business documents through one API
Every document below runs through the same authentication and comes back in the same response shape.
Identity documents. Aadhaar card, PAN card, passport, voter ID and driving licence.
Banking and income documents. Bank statements, passbooks, cheques and cancelled cheques, payslips (salary slips), income tax returns and credit cards.
Business and statutory documents. GST certificates, Udyam certificates, certificates of incorporation, shop establishment certificates, invoices and balance sheets.
Lending and legal documents. Loan agreements, legal documents and KYC forms.
Insurance, investments and other documents. Insurance policies, mutual fund statements, pension certificates, death certificates and utility bills.
Utilities that sit alongside them. Document identification, document quality check and document to text.
How Smart OCR Works
From capture to clean data
Four steps, and the first one is the one most teams skip.
- 01Check the captureThe quality check looks at blur, glare, lighting, cropping and resolution, and recommends a retake when an image won't read well. Catching a bad photo while the customer is still holding the phone costs seconds. Catching it in a review queue costs a callback.
- 02Identify the documentSmart OCR works out what the file is, and which side of it you're looking at where that applies, so a mislabelled upload reaches the right model instead of failing quietly.
- 03Read the fieldsThe model for that document pulls out the named fields, with confidence scores and quality or tamper flags on the documents that support them.
- 04Hand it to your systemThe JSON goes into your onboarding form, your loan origination system or the next API in the chain. In most flows that next call is a verification, because reading a document and trusting it are two different steps.
Recommend a retake. Tilt the card away from the light so the number is readable, then resubmit. No extraction has run.
Where Teams Use It
Four places Smart OCR earns its keep
- Digital onboardingA customer photographs an identity document and the form fills itself in, so nobody types a twelve-digit number on a phone screen.
- Loan file processingBank statements, payslips and income tax returns are read straight into your loan origination system instead of being keyed in by hand. When the credit team needs an assessment rather than the raw figures, the same files can go on to Document Analyzer.
- Business onboardingGST, Udyam and incorporation certificates are read into your business verification checks, so a company's details arrive already structured.
- Back-office digitisationLoan agreements, legal documents and KYC forms become searchable, structured records instead of scans in a shared drive.
Accuracy
What 95%+ means in practice
Smart OCR is built for 95%+ accuracy. What you see on your own files will move with the document type and the way it was captured.
A PAN card photographed flat in daylight is an easy read. A laminated Aadhaar shot under a branch tube light is harder, and a bank statement that has been photocopied twice is harder still. Regional-script fields, faded print and flash glare all pull the number down.
That is why the quality check is its own endpoint, and why sandbox access exists: run your actual mix of documents through it and measure what matters to your review queue. For more on what a misread field really costs, see the blog on the hidden cost of OCR errors.
Which is why the quality check is its own endpoint. Measure the number on your own captures, not on ours.
Where Your Documents Go
Hosted in India
Smart OCR runs on infrastructure hosted in India. In DPDP terms, your institution remains the data fiduciary and FinHub acts as a data processor on your behalf, so decisions about notice and consent stay with you.
If you need Aadhaar numbers redacted before documents are stored, that is a separate job handled by the Aadhaar Masking API, which you can run alongside Smart OCR.
Under DPDP
Your institution
Data fiduciary
Decides notice and consent
FinHub
Data processor
Processes on your instruction
Redacting the Aadhaar number before you store the file is a separate call to the Aadhaar Masking API, not part of this response.
Reference
Reference: What Each Endpoint Returns
The endpoint and the fields that come back, so you can size the integration before you write any code.
Identity documents
| Document | Endpoint | Key fields returned |
|---|---|---|
| Aadhaar card | POST /v1/ocr/aadhaar | Aadhaar number · Name · Date of birth · Address · Document quality flags |
| PAN card | POST /v1/ocr/pan | PAN number · Cardholder name · Date of birth · Document quality score · Tamper/quality flags |
| Passport | POST /v1/ocr/passport | Passport number · Name and gender · Date of birth · Address · Expiry date and place of birth |
| Voter ID | POST /v1/ocr/voter-id | EPIC number · Name · Date of birth · Document quality score · Tamper/quality flags |
| Driving licence | POST /v1/ocr/driving-license | Licence number · Name · Date of birth · Expiry date · Document quality flags |
Banking and income documents
| Document | Endpoint | Key fields returned |
|---|---|---|
| Bank statement | POST /v1/ocr/bank-statement | Account holder and account number · IFSC and bank name · Transaction date, description and amount · Running balance · Document quality flags |
| Passbook | POST /v1/ocr/passbook | Account holder name · Account number and IFSC · Bank and branch name · Latest balance and transaction date · Document quality score |
| Bank cheque | POST /v1/ocr/bank-cheque | Account number · IFSC code and bank name · Payee name and cheque number · Document quality flags |
| Cancelled cheque | POST /v1/ocr/cancelled-cheque | Account number · IFSC code and bank name · Account holder name (if printed) · Document quality flags |
| Payslip (salary slip) | POST /v1/ocr/salary-slip | Employer name and employee ID · Earnings breakdown · Deductions breakdown · Net pay and pay period · Document quality flags |
| Income tax return | POST /v1/ocr/itr | PAN and filer name · Assessment year · Gross total income · Tax paid and refund details · Document quality flags |
| Credit card | POST /v1/ocr/credit-card | Card number · Cardholder name · Expiry date · Issuing network · Field-level confidence scores |
Business and statutory documents
| Document | Endpoint | Key fields returned |
|---|---|---|
| GST certificate | POST /v1/ocr/gst-certificate | GSTIN · Legal and trade name · Business constitution · Registration date · Principal place of business |
| Udyam certificate | POST /v1/ocr/udyam-certificate | Udyam Registration Number (URN) · Enterprise name and type · Enterprise category (micro/small/medium) · Major activity · Registration date |
| Certificate of incorporation | POST /v1/ocr/incorporation-certificate | Corporate Identification Number (CIN) · Registered company name · Company type and category · Date of incorporation · Registrar of Companies (RoC) jurisdiction |
| Shop establishment certificate | POST /v1/ocr/shop-establishment | Business name and registration number · Registered address · Issuing authority and state · Validity or renewal date · Document quality score |
| Invoice | POST /v1/ocr/invoice | Vendor name and GSTIN · Invoice number and date · Line items and total amount · Tax breakup (CGST/SGST/IGST) · Document quality score |
| Balance sheet | POST /v1/ocr/balance-sheet | Total assets and liabilities · Equity and reserves · Key line-item breakdown · Reporting period and currency · Document quality score |
Lending and legal documents
| Document | Endpoint | Key fields returned |
|---|---|---|
| Loan agreement | POST /v1/ocr/loan-agreement | Borrower and lender name · Loan amount and currency · Tenure and interest rate · Repayment schedule · Field-level confidence scores |
| Legal document | POST /v1/ocr/legal-document | Full extracted text · Parties and signatories · Key dates and clause references · Document type classification · Processing status (sync/async) |
| KYC form | POST /v1/ocr/kyc-form | Customer name and address · PAN, Aadhaar or other ID numbers entered · Date of birth and contact details · Signature and form completeness status · Document quality score |
Insurance, investments and other documents
| Document | Endpoint | Key fields returned |
|---|---|---|
| Insurance policy | POST /v1/ocr/insurance-policy | Insurer name and policy number · Policyholder name · Sum assured and premium amount · Policy start and expiry dates · Document quality score |
| Mutual fund statement | POST /v1/ocr/mutual-fund | Fund house and scheme name · Folio number and investor name · Unit balance and NAV · Current market value · Document quality score |
| Pension certificate | POST /v1/ocr/pension-certificate | Pensioner name and PPO number · Sanctioning authority · Pension amount and category · Date of sanction · Document quality score |
| Death certificate | POST /v1/ocr/death-certificate | Deceased name and registration number · Date and place of death · Issuing municipal or registrar authority · Date of registration · Document quality score |
| Utility bill | POST /v1/ocr/utility-bill | Customer name and address · Service provider and account number · Bill date and due date · Billed amount · Document quality score |
Any document, and the utilities
| Document | Endpoint | Key fields returned |
|---|---|---|
| Any supported document (auto-classify) | POST /v1/documents/ocr | Structured fields (name, DOB, number, address…) · Document classification · Confidence scores · Tamper / authenticity signals |
| Document to text | POST /v1/ocr/document-to-text | Full extracted text · Per-line/word confidence scores · Detected language · Layout metadata |
| Document quality check | POST /v1/ocr/document-quality-check | Overall quality score · Blur, glare and lighting flags · Crop / edge-detection status · Resolution assessment · Retake recommendation |
| Document identification | POST /v1/ocr/document-identification | Detected document type · Classification confidence score · Side detection (front/back, where applicable) · Unsupported/ambiguous flag |
Compare
Smart OCR vs a General OCR Engine
| General OCR engine | FinHub Smart OCR | |
|---|---|---|
| What you get back | A block of text to parse yourself | Named fields as JSON |
| Document awareness | Doesn't know what it is reading | Identifies the document and returns the fields that belong to it |
| Indian layouts and scripts | You tune and maintain it | Built for Indian document layouts and regional script fields |
| Capture quality | Whatever you send is what you get | Quality check flags blur, glare and cropping before extraction |
| Tamper signals | Not part of the tool | Tamper and quality flags on supported documents |
| Template upkeep | Your team, every time an issuer redesigns a form | Ours |
Terms
Terms Worth Getting Right
The distinctions that decide whether a flow stands up under examination.
- OVD (officially valid document)
- The identity and address documents a regulated entity may accept for KYC, such as a passport, driving licence or voter ID. Which documents qualify for a given product is a question for your compliance team.
- Confidence score
- The model's own estimate of how sure it is about a field. Teams use it to decide which extractions to accept automatically and which to send to a person.
- Document classification
- Working out what a file is before reading it. It keeps a payslip from being run through a bank statement model, and it routes uploads by what they actually are rather than by the box the customer ticked.
FAQ
Smart OCR, Common Questions
Smart OCR is FinHub's AI-based OCR API for Indian financial services. It reads identity, banking, income and business documents and returns the fields on them as structured JSON, with quality and tamper signals on supported documents.
Related categories
Nearby on the API Platform
Last updated · 12 October 2026
Try Smart OCR on your own documents
Sandbox access with your real document mix, so you can measure accuracy on your captures rather than ours.