Tesseract, SimpleOCR and ABBYY FineReader run on your own computer. Only you see the page, there's no cost per page, and no internet connection is needed.
On-premise OCR, cloud only when it pays
SimpleIndex reads every page on your own hardware first, with no per-page cost and nothing leaving the building. When a field is too hard for local OCR, you choose whether it goes to a cloud service or to a model on your own server, and only those pages are sent. We used to call this Sunshine Software: software you own, running where your documents already are.
01
Three places a page can be read
One job can use all three. The local engines read every page; the other two only see what you send them.
AWS Textract, Azure, Google, and Claude, ChatGPT or Gemini. The provider reads the pages you send and bills you directly, per page or by tokens, with no markup from us. Needs an internet connection.
A language model running on your own hardware through Ollama, LM Studio or vLLM. Only you see the page, there's no per-page bill, and it works with no internet connection.
How a page decides where it goes
Local OCR reads every page
The page is made searchable on your own hardware, whatever happens next.
Zones, patterns and barcodes fill most fields
At no cost per page. On most documents, this is where the job ends.
Only the empty fields go further
To the cloud engine or the local model you picked, and only on the pages you allow.
Nothing is paid for twice
Fields the local engines read are never sent, and a field the model filled isn't asked for again.
03
What it costs over five years
Take 5,000 invoice pages a month, read by Textract's invoice model at AWS's published price of $10 per 1,000 pages (US West, as of October 2026).
Every page to the cloud: $50 a month, or $3,000 in AWS fees over five years, before the cost of the software that sends them.
SimpleIndex, local first: if local OCR finishes nine pages in ten, Textract sees 500 a month, or $300 over five years, plus a $1,800 Professional license paid once.
SimpleIndex with a model on your own server: the $1,800 license and the server it runs on, with no per-page fees at all.
Optional annual maintenance for support and upgrades isn't included, and the license keeps working without it. Check current AWS prices for your region. The full method, across five scenarios and three volumes, is in our total cost of ownership study.
04
Keeping pages in the building
Patient, client, customer and student records come with rules about who may see them. SimpleIndex runs recognition on your own hardware by default, so no outside service reads a page unless you send it one. If you use a cloud engine, the pages you send are processed by that provider, so check your agreement with them first.
How this plays out for healthcare, legal, financial services and education. It also runs with no internet connection at all, on air-gapped networks or at sites where the connection can't be trusted.
Before you decide where pages go
What is on-premise OCR?
OCR software installed and run on your own computers or servers, so documents are read where they already are instead of being uploaded to a cloud service.
Can SimpleIndex run with no internet connection?
Yes. The local engines and a model on your own server both work on air-gapped networks. Only the cloud engines need a connection.
Do I have to choose between local and cloud?
No. Each job decides, field by field. Most fields are read locally, and only the ones you choose go to a cloud engine or a local model.
What does a cloud engine cost?
The provider bills you directly: AWS Textract per page, the language model providers by tokens. Simple Software adds nothing per page. See Textract prices.
Is SimpleIndex licensed once?
Yes. A workstation license is paid once and keeps working. Annual maintenance, for support and new versions, is optional.
What was Sunshine Software?
Our name for software you own and run on your own hardware, rather than rent as a service. The idea hasn't changed; the cloud is now one of the engines it can use, not the alternative to it.
Which edition do I need?
Standard reads locally with Tesseract and SimpleOCR. Professional adds ABBYY FineReader, the cloud engines and language models, including one on your own server.
Try it on your own hardware
The free trial runs the full software on your own computer, local engines and cloud options included, and our team in Knoxville can help you set up the first job.
