# How much does OCR cost?

OCR is priced by the page, by a software licence, or by a service quote. Paxa OCR Lite costs 6.5 credits per page. See vendor rates and worked examples.

Published 2026-11-17. Paxa Labs.

OCR is usually priced by the page. Paxa OCR Lite costs 6.5 credits per page, which is $6.50 per 1,000 pages, whatever is printed on the page. A new account's 100 free credits read 15 pages with no card. A desktop program is sold by licence, and an outsourced service quotes by the job.

## What are the ways to buy OCR?

| Way to buy | How it is priced | Fits |
| --- | --- | --- |
| A desktop or online OCR program | A licence per user or a monthly plan, often with a page allowance | One person converting documents by hand, a few at a time |
| An outsourced scanning or data entry service | A quote per job after the provider sees sample pages | Boxes of paper that still need scanning, or work a person must check |
| A per-page OCR API or web console | A published rate per page, charged only for the pages read | Steady volumes, and documents that feed accounting or search systems |

A program or a service can run the same kind of reader that an API sells by the page. The difference is who does the work around it: scanning the paper, checking the result, and typing it into the next system.

## What decides the price of OCR?

| Factor | How it moves the price |
| --- | --- |
| Pages | The main unit. A photo counts as one page, and a PDF is counted before it is read. |
| Text or fields | Reading a page into text is the base rate. Pulling named fields such as a total or a tax ID is a higher tier at most vendors. |
| Tables and layout | Some vendors charge more to keep tables and reading order. Others include them in one rate. |
| Handwriting | Print reads more reliably than handwriting in any OCR tool. A service that checks handwritten pages by hand prices that checking. |
| Turnaround | A per-page API returns each document as it is sent. A service quotes a delivery date, and a rush costs more. |

## What do OCR vendors charge per 1,000 pages?

These are published list prices as read in October 2026. A range runs from a vendor's plainest reading tier to its most capable one. Its low end and its high end are different products. Paxa OCR Lite has one rate that returns Markdown with tables kept.

| Vendor | Reading text, per 1,000 pages | Extracting fields, per 1,000 pages |
| --- | --- | --- |
| Paxa OCR Lite and Paxa Document Extraction | $6.50 | $13 |
| LlamaParse | $1.25-56 | $10-75 |
| Amazon Textract | $1.50-4.00 | $15-70 |
| Azure Document Intelligence | $1.50-10 | $10-30 |
| Mistral OCR | $2-4 | $3-5 |
| Typhoon OCR | No published rate | No published rate |

Several vendors list a lower reading rate than ours on their plainest tier. Check which tier reads your documents before comparing, because each tier returns a different shape of output. [The OCR API comparison](https://paxalabs.com/compare/ocr-apis) links each vendor's price page and what its documentation says about languages. [Typhoon's FAQ](https://docs.opentyphoon.ai/en/faq) calls its API a free research showcase in beta, and [the Typhoon OCR comparison](https://paxalabs.com/compare/typhoon-ocr) covers its limits.

## What does Paxa OCR Lite cost for real jobs?

| Job | Pages | Credits | Cost |
| --- | --- | --- | --- |
| One receipt | 1 | 6.5 | $0.0065 |
| A contract | 20 | 130 | $0.13 |
| A month of single-page invoices | 1,000 | 6,500 | $6.50 |
| A paper archive | 50,000 | 325,000 | $325.00 |

One upload takes a file of up to 50 pages and 10 MB. The receipt and the contract are one upload each. The archive is many uploads, and a team with that volume usually sends files through the API at the same per-page rate. The page count and the charge appear before anything is read.

## Does extracting fields cost more than reading text?

Yes. Reading returns the whole page as text. Extraction returns only the fields you name, such as the invoice number, the date, and the total, each typed and ready for a spreadsheet. Paxa Document Extraction costs 13 credits per page for a field list of up to 50 fields, which is $13 per 1,000 pages. A longer list costs 19.5 credits per page, or $20 per 1,000. Reading the page is included, and one request takes up to 20 pages.

## What does invoice OCR cost?

An invoice or a tax invoice is usually one page. The cost follows the count. Reading 1,000 invoices into text costs $6.50. Pulling their fields with Paxa Document Extraction costs $13.00. The fields most teams pull are the seller's name and tax ID, the invoice number and date, each line item, the VAT amount, and the total.

[The invoice page](https://paxalabs.com/ocr/invoice) shows what makes a Thai invoice hard to read, and [Thai OCR for invoices and receipts](https://paxalabs.com/use-cases/invoices) covers the accounting use case.

## How many pages are free?

A new account gets 100 free credits with no card. They read 15 pages with Paxa OCR Lite, or fill fields on 7 pages with Paxa Document Extraction. Free credits are for personal use and evaluation. Buy any plan or credit pack, and the account's results are yours to use commercially, including those made before the purchase. [The Terms](https://paxalabs.com/terms#content-and-output) set this out.

## How do you read a first document?

1. Drop in a file: Open [the Documents tool](https://paxalabs.com/app/documents) and drop in a file. It accepts PDF, PNG, JPEG, WEBP, up to 50 pages and 10 MB. The page count and the cost appear before anything is read.
2. Pick the output: Choose Markdown to read the result yourself, or Structured to pass it to a program.
3. Check the numbers: Run it, then compare the amounts, dates, and ID numbers against the page.
4. Pull the fields: For an invoice or a receipt, open [the Extract tool](https://paxalabs.com/app/extract) and start from a ready-made field list.

## What should you read next?

- See a Thai tax invoice read field by field on [the OCR page](https://paxalabs.com/ocr), and filled into a schema on [the Document Extraction page](https://paxalabs.com/document-extraction).
- Turn documents into text for a language model with [PDF to Markdown](https://paxalabs.com/pdf-to-markdown), or into typed fields with [PDF to JSON](https://paxalabs.com/pdf-to-json).
- See every plan and credit pack on [the pricing page](https://paxalabs.com/pricing).
- Read [what OCR is](https://paxalabs.com/guides/what-is-ocr), [which Thai OCR fits each job](https://paxalabs.com/guides/best-thai-ocr), and [how to extract text from Thai PDFs](https://paxalabs.com/guides/thai-pdf-text-extraction).

## Common questions

### Is there free OCR?

Yes, within limits. A new Paxa account reads 15 pages on its free credits. Typhoon's FAQ calls its API a free research showcase in beta. Free tiers suit trying a tool on your own pages before a workflow depends on it.

### How is OCR charged?

Per page at most API vendors. Paxa OCR Lite charges 6.5 credits for every page, counted from the file before it is read. A blank page in a PDF still counts as a page.

### Does handwriting cost more?

Paxa OCR Lite charges the same rate for every page. Print reads more reliably than handwriting in any OCR tool. Try a few handwritten pages of your own before a workflow depends on them.

### Does the language change the price?

No. Paxa OCR Lite reads Thai, English, Simplified Chinese, and other languages at one rate, including pages that mix them.

### When is an outsourced OCR service worth paying for?

When the documents are still on paper and nobody on the team can scan them, or when a person must check every value. Ask whether the quote covers scanning, checking, and the file format you need back.
