# Run GOT-OCR 2.0

Reads scanned documents, including tables, forms, and handwriting. Ready in about 90 seconds on an RTX 4090, from $0.45/hr. No install, no code.

## What it does

GOT-OCR 2.0 turns a picture of a document into text you can search, copy, and put in a spreadsheet.

It handles the things that break older scanning tools: multi-column layouts, tables that must stay tables, mathematical notation, and reasonable handwriting.

The common jobs are invoices, receipts, contracts, historical records, and any archive that exists only as scans.

## Your setup

We run GOT-OCR 2.0 on an RTX 4090 with 24GB of memory, at $0.45 per hour, billed by the minute. It takes about 2 seconds per page.

The model itself is only 580 million parameters and needs about 8GB. The RTX 4090 is here for throughput rather than capacity: it processes pages in parallel batches, and on a large archive that parallelism cuts total time by roughly two thirds compared with the cheapest card, for less total cost.

## What it costs

5,000 pages takes about 2 hours 5 minutes of GPU time, which is about $0.94.

## Two ways to run it

- **Open the app.** Upload a document and see the extracted text beside the original.
- **Run a batch job.** Point it at an archive. Collect text, or a spreadsheet if the documents are structured.

## Model facts

| Field | Value |
|---|---|
| Parameters | 580M |
| License | Apache 2.0 |
| Memory required | 8GB minimum |
| Base model | Trained from scratch |
| Publisher | StepFun |
| Recommended hardware | RTX 4090, 24GB |
| Category | [Data processing](https://gpuvault.io/working-on/data-processing/) |

## Questions people ask

### Do I need to install anything to use GOT-OCR 2.0?

No. We start a machine with GOT-OCR 2.0 already loaded and hand you a link. Everything runs in your browser, there is nothing to download, and nothing is left on your computer afterwards. It works the same on a Mac, a Windows laptop, or a Chromebook. A workspace is usually ready in about 90 seconds.

### How much does it cost to run 5,000 pages?

About $0.94. GOT-OCR 2.0 takes about 2 seconds per page on the RTX 4090 we recommend, so 5,000 pages is roughly 2 hours 5 minutes of GPU time at $0.45 per hour. Billing is by the minute, and a batch job shuts the GPU off the moment the last item finishes. The calculator above works this out for your own numbers.

### Can I run GOT-OCR 2.0 on a cheaper card?

Sometimes, and the calculator will not always make it look worth it. The model itself is only 580 million parameters and needs about 8GB. The RTX 4090 is here for throughput rather than capacity: it processes pages in parallel batches, and on a large archive that parallelism cuts total time by roughly two thirds compared with the cheapest card, for less total cost. If you want to try a different card anyway, the advanced catalog lets you pick one and shows the estimated time before you commit.

### Can I use GOT-OCR 2.0 commercially?

Yes. GOT-OCR 2.0 is released under the Apache 2.0 license, which permits commercial use with no conditions worth worrying about. You can use the output in paid client work and in products you sell.

### What happens to my files after I stop?

Anything you save into your workspace folder stays in your account and is there when you come back. Everything else is destroyed: the machine is terminated and its disk is wiped. Batch job inputs and outputs are deleted from our storage 24 hours after you download them. We do not read your files or use them to train anything.

### What if it fails partway through?

You are not charged for work that did not complete, and failed sessions are refunded automatically. Batch jobs write each output as it finishes rather than at the end, so a failure at item 380 of 400 leaves you 379 usable files and a message saying what broke. Restarting picks up where it stopped rather than redoing the work.

---

*Source: https://gpuvault.io/model/got-ocr-2/ — GPUVault, GPU rental by the minute.*