What is the expected Textract performance (sync mode)?

Question

We are evaluating Textract against other solutions.

Extracting text from a one-page document with Textract takes a few seconds (**3-7 seconds**, depending on whether we use the layout feature or not). Is this the expected performance?

The same page gets OCRed by Google Cloud Vision API in 0.5 - 1.5 seconds. I wonder if we can achieve comparable performance with Textract.

Currently on a free tier.
Using synchronous API.
Region - us-east-2.

Code is like this (uses wrapper library, but got similar results using boto3):

```
from textractor import Textractor
extractor = Textractor(profile_name="default")
document = extractor.detect_document_text("document.png")
```

Thank you!

Answer

Hi Andrey, thanks for using AWS Textract, to help you better with performance we would like to take a look at kind of documents being used to call Textract. You can cut us a support ticket and share document there (https://docs.aws.amazon.com/awssupport/latest/user/case-management.html#creating-a-support-case), however if you have Basic Support on account, you might not be able to create a support case. In that case, I'd recommend depending on nature of document, you can share it via any publicly available file system.

What is the expected Textract performance (sync mode)?

Relevant content