Available in Classic and VPC
You can have your questions answered quickly by referring to the answers in the FAQs before reading the user guides. If your questions are not resolved in the FAQs, see user guides to find the information you want.
CLOVA OCR FAQ includes the following topics:
General
Find answers to common questions about CLOVA OCR below.
Q. What are the key features of CLOVA OCR?
A. Powered by NAVER’s AI technology, CLOVA OCR uses high-performance OCR recognition models optimized for core business use cases. It accurately recognizes documents and improves document processing efficiency by allowing you to define follow-up workflows for faster processing. It also uses proprietary text area detection and recognition technologies to understand various text formats. In addition, the service lets you easily create templates, designate specific areas, and quickly extract only the text you need.
Q. Can document processing be automated?
A. By using CLOVA OCR, documents that previously required manual sorting can be automatically classified based on their similarity to registered templates. This enables document classification without user intervention and helps you design effective workflows.
Q. What is the difference between CLOVA eKYC and CLOVA OCR?
A. CLOVA OCR analyzes various user-uploaded images and extracts features in the format you need. In contrast, CLOVA eKYC uses Document OCR in CLOVA OCR to extract only features from IDs and business registration certificates, and verifies their authenticity through third-party certification authorities.
Q. Can CLOVA OCR be configured to recognize English only?
A. Recognizing English exclusively is currently not supported.
Q. Can CLOVA OCR validate text values after recognition?
A. Yes. CLOVA OCR can validate recognized text values, helping reduce repetitive validation tasks and improve reliability. It can also be integrated with services such as Papago, helping you provide more tailored services to users.
Q. What are the criteria for determining linebreak values in General OCR?
A. The linebreak value indicates line-level information for text recognized in an image. It shows whether the text is at the end of a line. If the text is at the end of a line, the value is returned as “true”.
As of April 23, 2026, the criteria for determining linebreak values for General OCR in the Korea Region have changed to be based on paragraphs and words as follows:
| Before | After |
|---|---|
 |
 |
API integration
Find answers to common questions about API integration below.
Q. How can I find the IP information required for firewall whitelisting to use CLOVA OCR within an internal network?
A. You can get the IP information by configuring a VM and routing calls through it. For more information on how to integrate VM PROXY, see CLOVA OCR VM PROXY integration guide in the CLOVA OCR API guide.
Q. I want to control access using the IP information of the API Gateway Invoke URL. How can I get the information required for access control?
A. You can get the access control information by configuring a VM and routing calls through it. For more information on how to integrate VM PROXY, see CLOVA OCR VM PROXY integration guide in the CLOVA OCR API guide.
Q. Using a SecretKey to call CLOVA OCR from a client may expose the SecretKey and cause security issues. Is there a way to use it securely without exposing the SecretKey?
A. Yes. You can prevent SecretKey exposure by configuring a VM and routing calls through it. For more information on how to integrate VM PROXY, see CLOVA OCR VM PROXY integration guide in the CLOVA OCR API guide.
Security
Find answers to common questions about security below.
Q. Is there any risk that images or documents sent to CLOVA OCR are stored separately or used for other purposes?
A. No. CLOVA OCR processes target data received through the API in the OCR service engine and returns recognized text results. Images and documents sent by users are not stored in the cloud, nor are they used to improve engine performance or analyze data patterns. Returned text results are not included in log data, and only OCR call access history is retained. In other words, the service does not store the original data recognized by OCR or the recognition results. It stores only the minimum log data required to verify billing based on OCR call access history.
Q. Does CLOVA OCR provide security features, such as encryption, for returned text results?
A. CLOVA OCR is integrated through API Gateway. When using CLOVA OCR, the connection between your server or application and the NAVER Cloud Platform API Gateway is encrypted via HTTPS. CLOVA OCR also processes original images in memory and returns the results without storing them separately. Recognition results are not stored either. This allows you to use the service securely.
Pricing
Find answers to common questions about pricing below.
Q. How are fees calculated when using the table extraction feature in CLOVA OCR?
A. If a table structure is recognized in the uploaded image, both text extraction and table extraction fees will apply. If no table is detected, you will only be charged for text extraction.