Check OCR
Automatically extract key information from bank checks using our OCR API
.svg){kind=icon}
Detect printed and handwritten fields
Extracts both machine-printed and handwritten data found on physical checks.
Identify all key check values
Automatically captures check number, routing/account codes, amounts (digits and letters), and issue date.
Adapt to different check layouts
Handles domestic, business, and international checks without needing layout-specific configuration.
Designed for software integration:
REST API with SDKs
Plug in using Python, JavaScript, or cURL—with complete examples and integration guides.
Sub-second response time
Get results instantly—ideal for interactive UIs and batch processing alike.
Normalized and structured outputs
Receive JSON with clearly labeled values, position data, and confidence scores.
Flexible deployment options
Use in the cloud or on-premise, depending on your regulatory and infrastructure needs.
Transform your workflow with:
Capture check visuals with a simple upload
Drag-and-drop, scan, or use a phone camera—Mindee processes them all.
Extract data without manual input
Everything from routing codes to issue dates is automatically parsed.
Validate for completeness
Flag missing fields or low-confidence results before submission.
Feed into downstream tools instantly
Send data to ledgers, compliance platforms, or dashboards—no manual steps required.
Cut delays and reduce errors in all stages of check intake and processing:
- Capture physical or scanned checks
- Extract all standard fields with no template setup
- Validate data for completeness and accuracy
- Transfer results directly to banking systems or ledgers
- Log results for audit or fraud detection workflows
- Support legacy or hybrid paper/digital workflows
Tailored for banking and fintech use cases
Mindee’s Check OCR is trained on financial documents and workflows.
Fast, accurate, and easy to deploy
Accuracy above 95%, results in < 1s, and quick setup with SDKs.
No templates or rules to maintain
Mindee’s pre-trained model adapts to layout variations out of the box.
Developer-first documentation
API references, Postman collections, and walkthroughs included.
Built-in confidence scoring
Helps flag risky or incomplete extractions for review.
GDPR & SOC 2 Type II compliant
Process sensitive data with confidence and full compliance.
check
Start automating your fraud workflow today. Get started with a free trial or schedule a demo to learn more.
Frequently Asked Questions
Common questions about our intelligent document processing solutions.
How can I test the bank check OCR API?
The bank check OCR API is in private beta. If you are interested in testing it, request an early access.
A member of our team will organize a demo and answer any questions you might have.
What is the pricing?
Start trying the check OCR API thanks to the free plan that includes 250 pages monthly.
For monthly consumption above 250 pages, the pricing starts at $0.10 per page and is decreasing gradually. Check our pricing page to see available plans or reach out to our team to know more.
Check OCR : What are the supported countries?
What are the supported countries?
Our technology is based on Computer Vision, which means that it relies on image analysis rather than text analysis to extract bank check data. As such, there is no language limitations to the check OCR API.
Does it support handwritten fields?
Handwritten fields such as amounts are extracted.
For signatures, the position of the signature on the check is detected and outputted in the API response. If there are multiple signatures on the document, the different positions will be included in the response.
Does the API support MICR characters?
The MICR Line uses special character to indicate routing number, account number and check number on the check. This 3 MICR codes are used for processes and clearance purposes.
Our models have been trained to support those characters thanks to a specialized MICR OCR. Those 3 fields are available in the API response.
Check OCR : Does the OCR work on low-quality image?
Does the OCR work on low-quality image?
Bank check OCR works on digital bank checks and also paper bank checks that are scanned. Our models are trained on a wide variety of documents to make our API robust to various image quality.
To take into account potential ink stains on the check or blur on the scan, we train our models including augmentations. The OCR will be able to read most of the documents as long as they are readable.
Check OCR : How difficult is it to integrate the API?
How difficult is it to integrate the API?
The integration is seamless and straight forward. There are different ways of integrating and accessing our API products.
We recommend using our client libraries available in Python, Node, Ruby and Java.
A custom integration can be made using our REST API.
Does Mindee offer technical support?
If you have any questions, feel free to reach out to our team as we offer a wide range of support:- Access our API and client libraries documentation
- For technical questions regarding integration get help from our slack community
- For pricing plan and commercial questions use our chat and meet with our team