DocView OmniRead: OCR and handwriting APIRecognition that stays on your servers
DocView OmniRead is a recognition engine and API that runs on your own servers. It reads printed text in 50 languages, handwriting and barcodes on ordinary processors, with no graphics card and no cloud service. $100 a month per server, unlimited pages.

- $100/monthPer server
- Unlimited pagesNo page limits
- 50 languagesPrinted text (OCR)
- CPU onlyNo graphics card, no cloud
One engine for every way to read a document
Use it to turn page images into text, handwriting values and barcode values, each with a confidence.
Printed text (OCR)
50 languages, with the position and confidence of every line and word. Fast and accurate modes.
Handwriting (ICR)
Hand printing and handwriting in zones or across the page, with a confidence for every value.
Barcodes and QR codes
Every common 1D and 2D type, at any angle, several per page, including small and damaged codes.
Vision model (optional)
Ask a question about a page or a zone, such as "What is the invoice number and the total?", or have it transcribed. Off by default.
Image clean-up
Straightening, noise removal, black-and-white conversion, contrast and form-line removal.
Commercially licensable
Built on commercially licensable models only. Models are checked before use and can be copied onto offline servers.
Inside DocView Capture, or behind your own applications
Inside DocView Capture
- It is the OCR engine of a Batch Setup, plus the handwriting and vision-model finders
- Several OmniRead services share the work, with failover
- Group services into pools so each Batch Setup uses only its own pool
- Status, settings and upgrades are managed from DocView Capture
Standalone OCR API
- Its own admin page, which works offline, with its own sign-in
- API keys per application and usage counts
- HTTP and JSON API with an OpenAPI 3 description
- Examples in curl, Python, C# and JavaScript, and a Try it page
Your documents never leave your servers
OmniRead runs on your own hardware. By default it sizes itself to the processor, keeps one or two cores free and reads at below-normal priority, so other programs on the server stay responsive.
- Windows service, installed by DocView Setup
- Linux container image
- Several named OmniRead instances can run on one machine
DocView OmniRead
$100 /month per server
Unlimited pages. OCR, ICR, barcodes and image clean-up.
- 30-day evaluation from first start
- Use inside DocView Capture is covered by the DocView Capture licence
DocView OmniRead questions
What does DocView OmniRead read?
Printed text (OCR) in 50 languages with the position and confidence of every line and word, handwriting (ICR) in zones or across the page, every common 1D and 2D barcode including QR codes, and, with the optional vision model, answers to plain-language questions about a page or zone.
Do I need a graphics card or a cloud service?
No. DocView OmniRead runs on ordinary processors (CPU only) on your own servers. No graphics card and no cloud service are needed. Models can be copied in for offline servers.
Can I use OmniRead from my own applications?
Yes. Standalone OmniRead has its own admin page, API keys per application, usage counts, a Try it page, and an HTTP and JSON API with an OpenAPI 3 description and examples in curl, Python, C# and JavaScript.
How much does DocView OmniRead cost?
$100 per month per server, with unlimited pages. Every new server runs a 30-day evaluation from first start. Use inside DocView Capture is covered by the DocView Capture licence.
Which platforms does it run on?
As a Windows service installed by DocView Setup, or as a Linux container image. Several named OmniRead instances can run on one machine.
Will it slow down other programs on the server?
By default OmniRead sizes itself to the processor, keeps one or two cores free and reads at below-normal priority so other programs stay responsive.
Read documents on your own hardware
Ask for an evaluation licence, or see OmniRead at work inside DocView Capture.


