The AI-enhanced OCR software platform launched by Tungsten Automation (formerly Kofax) provides three delivery forms: desktop (Standard/Ultimate), server (Server) and embedded SDK (Capture SDK). It supports document recognition, layout maintenance, barcode recognition and automated workflow orchestration in 120+ languages.

OmniPage AI: Tungsten Automation’s enterprise-grade AI-enhanced OCR platform

Core parameters and statistics

Specific technical parameters (such as model size, context length, supported file formats, input and output restrictions, etc.) are subject to the official product page. It is recommended that users verify the latest technical specifications and system requirements before choosing to ensure that they match their own usage scenarios.

User and market recognition

Gradually build user awareness in the field, and product capabilities are used by content creators and teams to improve work efficiency. Some industry users have incorporated it into their daily workflow. It is recommended to refer to the latest official disclosures for specific user scale and industry adoption rate data.

Cost advantage

C-side/individual user cost

Product	Price (SGD)	Delivery Method	Applicable Scale
OmniPage Standard 18.0	$149 SGD	Buyout	Light/Infrequent Document Conversions
OmniPage Ultimate 19.2	$149 SGD	Buyout	Medium and high-frequency document processing
Ultimate Free Trial	Free	Trial limited	<1000 pages/week

Desktop pricing presents an interesting feature: Standard is priced the same as Ultimate ($149 SGD), which means users only have to consider feature differences rather than price thresholds. Ultimate leads the way in batch processing of PDF searches, advanced image preprocessing and other dimensions, making it a more cost-effective choice for most users.

API/Developer Cost

Products	Pricing Model	Applicable Scenarios
Capture SDK for Windows	Business inquiry required	Embed OCR capabilities in Windows applications
Capture SDK for Linux	Business inquiry required	Integrate OCR recognition pipeline on Linux server
Capture SDK for Mac	Business inquiry required	Integrate document capture capabilities in macOS applications

SDK pricing is not disclosed, please contact Tungsten Sales for a quote. Key variables that affect pricing include: the number of deployment nodes, the annual volume of documents processed, and whether source code level customization is required. It is recommended to apply for an evaluation license from the manufacturer during the PoC stage and confirm that the technical specifications are met before entering into business negotiations.

Business / Privatization Costs

Products	Pricing Model	Applicable Scenarios
OmniPage Server	Business inquiry required	Internal high-concurrency document conversion service
Volume License	Tiered Pricing	Organizations with Multiple Desktop Deployments

The core cost of enterprise-level procurement is not the software itself, but integration and maintenance. OmniPage Server requires Windows Server context to run, and the operation and maintenance team needs to have basic Windows service management and performance tuning capabilities. For scenarios with an annual processing volume of more than 10 million, it is recommended to simultaneously evaluate the Total Cost of Ownership (TCO), including license, hardware, operation and maintenance, and training costs.

Hidden cost reminder

• Windows Locked: The desktop version only supports Windows systems, Mac/Linux users need to use it indirectly through a virtual machine or SDK.
• Upgrade Rhythm: Major version upgrades of buyout products usually require additional payment, and long-term holding requires an upgrade budget.
• Learning Curve: Ultimate has a high function density, and some advanced features (such as custom workflow Zone OCR) require a certain amount of training investment.

Main functions

1. AI enhanced optical character recognition (AI-OCR)

OmniPage's core capability is to convert text in scanned PDFs and images into editable, searchable text. Its AI engine superimposes a computer vision pre-processing module on the basis of traditional OCR, including adaptive binarization, automatic rotation/correction, noise removal and resolution optimization, which greatly improves the recognition rate of low-quality scans and documents captured by mobile phones.

• Actual value: Enterprises can digitize piles of paper files in batches without manual pre-processing. Contact information can be extracted from a casual photo of a business card, and a tilted scan of an invoice can be restored with high accuracy.

2. Layout Analysis and Restoration (Layout Analysis)

The engine automatically detects columns, tables, charts, and text areas in the document and preserves the original layout to the greatest extent possible when outputting. This capability is particularly critical in the conversion of complex formatted documents such as technical manuals, academic papers, and multi-column press releases.

• Actual value: When converting from PDF to Word, paragraph instances, table positions, headers and footers do not fly around, and editors do not need to adjust the format page by page.

3. Barcode & ICR

It supports common one-dimensional code (Code 39, EAN, UPC) and two-dimensional code (QR Code, Data Matrix) recognition, and also provides handwritten character recognition (ICR) capabilities.

• Actual value: In logistics and manufacturing, barcode scanning and OCR can be processed in one; in hospital scenarios, handwritten prescription orders can be partially automated through ICR.

4. Batch processing and automated workflow

The Ultimate version provides a batch processing scheduler that supports automatically grabbing files from folders or emails, queuing conversions, and routing output according to preset rules. Users can customize the "one-click process" - compressing repeated operations of multiple clicks into one button.

• Actual value: The finance department can completely automate the scanning, identification, archiving and exporting process of hundreds of invoices every week, and the personnel change from "operators" to "auditors".

5. Multi-format output and PDF in-depth processing

Supports output to PDF (including searchable PDF), Microsoft Office (Word, Excel, PowerPoint), Corel WordPerfect, HTML, and ePub. The Ultimate version comes with eDiscovery Assistant, which can convert single or batch PDFs into fully searchable document libraries.

• Real Value: Legal teams can batch convert thousands of pages of case materials into searchable PDFs, reducing keyword lookup time from days to minutes.

[Expert View] Hidden synergies between functions

Each function of OmniPage is not an isolated module, but forms a complete chain of "input → preprocessing → recognition → structuring → output → routing". It makes no sense to look at the OCR accuracy figures alone - what really reduces the user's workload is the batch scheduler's serial ability to pull up automatic preprocessing → multi-engine parallel recognition → format fidelity output → automatic archiving according to rules. This is also the core gap that distinguishes OmniPage from free online OCR tools: it is not a single point recognizer, but a document processing pipeline.

Model and version evolution

OmniPage's product line evolved through three main phases.

Phase One: Nuance/Caere Era (1990s–2010s)

OmniPage was originally developed by Caere Corporation and later acquired by ScanSoft (later renamed Nuance Communications) in 2002, becoming an important part of Nuance's document product line. Versions during this period focused on basic OCR accuracy improvements and Microsoft Office output compatibility, with version numbers iterating from the early OmniPage Pro to OmniPage 18/19.

Phase 2: The Kofax Era (2019–2024)

In 2019, Nuance's document imaging division was acquired by Kofax, and OmniPage was included in the Kofax product matrix. Kofax will release OmniPage Ultimate 19.0 around 2022, introducing an AI-enhanced image pre-processing engine for the first time to improve the recognition success rate of documents captured by mobile phones to close to the level of scanners. Version 19.1 further optimizes table recognition and multi-column layout maintenance.

The third phase: Tungsten Automation era (2024–present)

In 2024, Kofax changed its name to Tungsten Automation and OmniPage continued to evolve as its core OCR product line. The latest version, Ultimate 19.2, continues to iterate in terms of PDF in-depth search eDiscovery functionality and 120+ language recognition.

Version	Time Window	Key Changes
Standard 18.0	~2023	The latest version of the Standard product line, supporting Windows 10/8/7/XP
Ultimate 19.0	~2024	Introducing AI image preprocessing engine to greatly improve low-quality document recognition rate
Ultimate 19.1	~2024	Optimized table recognition, multi-column document retention, new ePub output
Ultimate 19.2	~2025	Enhanced PDF search with eDiscovery Assistant, optimized digital camera image OCR

Version Selection Suggestions: Standard is suitable for occasional document conversion needs and has a streamlined feature set; Ultimate covers complete capabilities from preprocessing to batch scheduling, and is the main choice for most organizations. Server and Capture SDK are designed for scenarios where OCR needs to be embedded into your own system or support extremely large batches, and their version numbers are not necessarily synchronized with the desktop version.

Technical advantages

Hybrid OCR engine architecture

OmniPage's core competitiveness lies in the fact that its OCR engine accumulated over many years is not a single algorithm, but a multi-engine hybrid architecture. The system runs a traditional rule-based OCR engine and a deep learning-based AI recognition engine at the same time. After generating multiple recognition hypotheses for the same area, it selects the optimal result through a confidence voting mechanism. This "dual-engine parallel → voting arbitration → result fusion" mechanism shows a more stable accuracy than a single engine solution when identifying documents with noise, low contrast or non-standard fonts.

Image preprocessing pipeline

OmniPage's AI preprocessing module automatically performs the following sequence of operations before recognition:

1. Adaptive Binarization - dynamically determine the threshold based on the local brightness characteristics of the image to solve the problem of uneven illumination
2. Automatic correction and rotation detection - identify document orientation and correct tilt (supports automatic correction up to ±15°)
3. Noise Filter - Remove salt and pepper noise and background texture during scanning
4. Resolution Optimization——Super-resolution enhancement of images with too low resolution

The value of this preprocessing pipeline is: reducing manual intervention. Operators do not need to manually adjust scanning parameters or pre-process images, the original document is fed "as is" to obtain usable recognition results.

Layout Preservation

OmniPage uses a "region-aware" algorithm to map the recognized text into the layout structure of the original document during the output stage. Unlike traditional OCR tools that produce confusing paragraphs and broken tables when outputting Word, OmniPage identifies column boundaries, table cells, and legend relationships on the page, and reconstructs an approximate visual structure in the target format.

• Key points of project implementation: A hybrid layout segmentation strategy based on connected domain analysis + top-down recursive segmentation. First identify large areas (text, images, tables), and then subdivide the text area by lines and paragraphs. This method is better than the end-to-end layout segmentation solution of pure deep learning on complex multi-column documents, because the rule part ensures the certainty and interpretability of the results.

Enterprise-grade scalability

The Server version supports horizontal scalability and can distribute identification tasks to multiple computing nodes through load balancing. Capture SDK allows developers to natively integrate OCR capabilities into C++, C# and Java applications, making it suitable for embedded specialized recognition scenarios.

Why OmniPage may be faster on the same hardware: Its engine has higher parallel utilization of multi-core CPUs (multiple pages can enter different stages of the pre-processing→recognition→post-processing pipeline at the same time), rather than being processed one by one by a single thread. This architectural design has significant throughput advantages over single-page processing tools (such as Tesseract's default configuration) in batch scenarios.

How to use

How to get it

Path	Entrance	Applicable objects
Ultimate free trial	Fill out the trial form on the official website	Personal user evaluation
Online purchase (buyout)	Order directly from the official website	Individual or enterprise stand-alone deployment
Business inquiry (Server/SDK)	Contact sales team	Enterprise batch deployment or embedded development
30-day money back guarantee	If you are not satisfied after purchasing, you can apply for a refund	Reduce purchasing risk

Basic steps for using the desktop version

1. Installation: Download the trial version from the official website or obtain the installation package after purchase, and run the installation wizard on the Windows system. The system needs to meet the minimum requirements: 1 GHz processor 1 GB RAM 2.7 GB disk space.
2. First time configuration: Start OmniPage and select a workflow template (such as "Scan to Word", "PDF to Searchable PDF", "Batch Convert Folders"). The Ultimate version provides a Launchpad wizard interface.
3. Document Input: Supports direct import from scanner (WIA/TWAIN/ISIS driver), local files (PDF, TIFF, JPEG, PNG, etc.) or digital camera.
4. Identification and Conversion: Click the "Recognition" button, and the engine automatically completes preprocessing → OCR → layout analysis → output generation. The entire process is completed locally, no network connection is required.
5. Export and Share: Select the output format (Word, PDF, Excel, HTML, ePub, etc.), and the file can be saved locally, emailed directly, or routed to a preset workflow.

Developer integration path

• Capture SDK for Windows: supports C++/C# integration, provides COM and .NET interfaces, and is suitable for embedded OCR in desktop applications.
• Capture SDK for Linux: supports C++ API, suitable for server-side background integration, and compatible with Docker containerized deployment.
• Capture SDK for Mac: The feature set is basically the same as the Windows version, and it is oriented to OCR integration of macOS native applications.
• OmniPage Server: Provides RESTful API, supports Java/C#/Python client calls, and is suitable for document recognition modules in microservice architecture.

Typical calling process (Server API mode)

File upload → Server queuing → AI preprocessing → OCR recognition → Structured output → Callback notification

Developers only need to focus on file submission and result reception, and the intermediate processing details are encapsulated by the server.

Product Pricing

Pricing Overview

OmniPage adopts the buyout system as its core pricing model, forming differentiated competition with the current mainstream AI SaaS subscription system. The specific price depends on the region. The following is the price reference in Singapore dollars (SGD):

Product	Selling price (SGD)	Nature of fee	Is tax included
OmniPage Standard 18.0	$149	One-time purchase	Excludes local taxes and fees
OmniPage Ultimate 19.2	$149	One-time purchase	Excludes local taxes and fees
OmniPage Server	Request a quote	Annual subscription or one-time	Subject to contract
Capture SDK	Request a quote	Charged per development node	Subject to contract
Volume License	Tiered discounts	One-time buyout	Excluding local taxes

Free trial and refund

• Ultimate offers a fully functional free trial limited to scanning no more than 1,000 pages per week.
• You can apply for a full refund within 30 days after purchasing the personal desktop version (need to purchase through the official website).
• For commercial volume licensing and Server/SDK trials, you need to contact the sales team to apply for an evaluation license separately.

Hidden cost list

1. Large version upgrade fee: Major version upgrades of buyout products (such as 19.x to 20.0) usually require the purchase of an upgrade license, and an upgrade budget needs to be reserved for long-term holding.
2. Hardware Investment: The Server version requires an independent Windows Server instance, and the memory and CPU configuration affect parallel throughput performance.
3. Training Cost: Advanced functions (Zone OCR, customized workflow) require a certain amount of learning investment. It is recommended to reserve training costs in the procurement budget.
4. Technical Support: Basic support for the desktop version is included in the purchase price, and enterprise-level technical support contracts (SLA) need to be purchased separately.

Application scenarios

1. Digitization and archiving of corporate documents

Scenario Description: Enterprises produce thousands to millions of paper documents every year - contracts, invoices, reports, personnel files. The traditional approach is to manually enter or scan into a non-searchable PDF.

Benefits of OmniPage: Batch scan → Automatic OCR → Classification by rules → Export to searchable PDF/Word → Archive to document management system. The processing density changes from "human flesh page by page" to "machine batch", and the digital efficiency of large document centers can be increased by 5-10 times (according to industry deductions, the specific values ​​vary depending on the scenario).

2. Invoice and financial document processing

Scenario description: The financial department receives a large number of invoices every week, with different formats and uneven quality. It needs to extract structured fields such as invoice number, date, amount, etc.

Benefits of OmniPage: Predefine the invoice recognition area through the Zone OCR function, and cooperate with the InvoiceAgility solution to realize the process from scanning to automatic entry. Barcode recognition can automatically associate purchase order numbers, reducing manual matching time. In this scenario, the boundary between human and machine collaboration is clear: the machine completes identification and field extraction, and financial personnel only need to verify abnormal items.

3. Digitization of medical records

Scenario Description: A hospital needs to convert paper medical records, test reports, and prescription orders into data that can be retrieved by an electronic health record (EHR) system.

OmniPage’s benefits: The ICR engine can partially recognize handwritten prescriptions, and the AI ​​preprocessing module improves the recognition quality of old medical records. 120+ language support is particularly important in medical scenarios in multilingual regions (such as Singapore and Hong Kong). This scenario requires manual confirmation: key data involving patient privacy (diagnostic conclusions, medication dosage) must be verified twice by medical staff.

4. Batch processing of legal documents

Scenario Description: Law firms or legal departments need to process thousands to hundreds of thousands of pages of documents for keyword search, classification and indexing in litigation, mergers and acquisitions and other projects.

OmniPage Benefits: eDiscovery Assistant converts large batches of PDFs into fully searchable document libraries that attorneys can perform keyword searches in minutes instead of flipping through paper documents page by page. Batch processing’s out-of-hours scheduling capabilities have immediate value for legal teams’ time-sensitive delivery.

5. Label identification in logistics and manufacturing

Scenario description: Warehouses and factories need to read data such as batch numbers, serial numbers, and expiration dates on product labels for inventory management and quality traceability.

Benefits of OmniPage: Barcode recognition and OCR are integrated into one process, and one scanning action can obtain text and barcode information at the same time. Capture SDK can be integrated into handheld devices or assembly line camera systems for real-time recognition.

Applicable people

Enterprise Document Management Team

Adaptation value: Direct users of the document digitization project. OmniPage's batch processing and workflow automation capabilities increase paper-to-digital conversion efficiency by an order of magnitude. It is important to note that the time distribution of document volume needs to be evaluated before deployment - throughput bottlenecks during peak periods may force the team to upgrade to the Server version.

Finance and Accounting Team

Adaptation value: The core beneficiary of automated invoice processing. Zone OCR can pre-configure invoice templates, and the batch scheduler can pull scan folders regularly. The premise of application is that the invoice format is relatively standardized - if a large number of handwritten or unstructured invoices are received, the accuracy of ICR will decrease, and manual verification will need to be supplemented.

IT and Development Team

Adaptation value: Developers who need to embed OCR capabilities in self-built systems or RPA pipelines. Capture SDK provides a platform-independent identification interface, and the server version of the REST API is suitable for microservice integration. Before getting started, you need to have some development experience in Windows (desktop SDK) or Linux (server SDK).

Legal and Compliance Team

Adaptation value: A direct beneficiary of the eDiscovery scenario, the batch PDF search capability significantly improves the efficiency of the document review stage. However, it should be noted that OmniPage does not provide NLP-level content understanding - it does "text recognition" rather than "semantic understanding", and advanced contract analysis still requires a dedicated eDiscovery platform.

Applicable boundaries and preconditions

• Not suitable for people: Teams who need online collaborative editing (OmniPage is a stand-alone desktop software, not a cloud collaboration platform); individual users who only need to occasionally convert a few pages of PDF (free online OCR tools may be more lightweight); AI engineers who want to obtain structured JSON directly after recognition (it is recommended to consider cloud OCR services such as Cloud Vision API).
• Hard prerequisites: The desktop version requires a Windows system; all product lines require a network connection for activation and updates (the recognition process itself can run offline); the Server version requires a Windows Server environment and does not support Linux native deployment (Linux users need to use the Linux version in the SDK).

Summary and Outlook

OmniPage AI's core competitiveness lies in three irreproducible accumulations: 30+ years of iterative accumulation of OCR engines, a full product line matrix covering desktops to servers, and the channels and integration ecosystem established by Tungsten Automation in the field of enterprise document automation. It's not the cheapest OCR tool on the market, but for organizations that take document processing seriously - bank back offices, hospital record rooms, law firm document centers - it's a time-tested, high-reliability choice.

Current limitations: The desktop version only supports the Windows platform and does not support cloud native deployment. The update pace of AI models is slower than that of cloud AI services (such as Google Cloud Vision or Azure AI Document Intelligence); it does not have semantic understanding capabilities and cannot automatically identify document types or extract unstructured fields; large version upgrades require additional fees.

Procurement and Adoption Risk Assessment: It is recommended that organizations proceed in the rhythm of "trial first (Ultimate free version) → small-scale PoC (select a high-frequency document type) → verify recognition rate and throughput indicators → expand to the Server version as needed." Before purchasing, enterprises need to focus on verification: whether the peak distribution of the current document volume is within the horizontal expansion capability of the Server version; the pricing terms and validity period of the upgrade license; and whether the response time of the technical support SLA meets business continuity requirements. For organizations with an annual processing volume of less than one million pages, the Ultimate desktop version combined with batch scheduling is sufficient; only extremely large-volume scenarios are worth initiating the purchase process of the Server version or Capture SDK.

Version Info

• OmniPage Ultimate 19.2 ：The latest version of Ultimate supports Windows 11/10/8/7/XP, optimizes the OCR accuracy of digital camera images, enhances PDF search and eDiscovery Assistant functions, and supports 120+ language recognition. （~2025）
• OmniPage Ultimate 19.0 ：Introducing an AI-enhanced image pre-processing engine to significantly improve the recognition rate of low-quality scans and documents captured on mobile phones, and a new batch processing scheduler. （~2024）
• OmniPage Ultimate 19.1 ：The table recognition and layout maintenance capabilities have been optimized, the parsing effect of complex multi-column documents has been enhanced, and the ePub output format has been added. （~2024）
• OmniPage Standard 18.0 ：The latest version of the Standard product line is designed for light document conversion needs, supports output formats such as Microsoft Office, Corel WordPerfect, and HTML, and is compatible with Windows 10/8/7/XP. （~2023）