> ## Documentation Index
> Fetch the complete documentation index at: https://sensible.mintlify.site/llms.txt
> Use this file to discover all available pages before exploring further.

# OCR engine

Specifies the optical character recognition (OCR) engine for extracting text from images. For information about additional OCR options, see [OCR](/senseml-reference/concepts/ocr).

### Enums

The following table shows the enums available for the OCR Engine parameter.

| enum      | description                                                                                                                                                                                                                |
| --------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| Amazon    | Default engine for the OCR preprocessor.                                                                                                                                                                                   |
| Microsoft | Default engine for document types.Suited to typewritten documents and large documents up to 50 MB in size.                                                                                                                 |
| Lazarus   | Faster than Microsoft and produces similar output.                                                                                                                                                                         |
| Google    | Faster than Microsoft and suited to handwriting and documents that are 5 pages or fewer. The Google engine doesn't merge words into lines automatically. Use the Merge Lines preprocessor in your configurations to do so. |

**Note:** When Sensible extracts from [portfolios](/layout-based-extractions/portfolio), it uses Microsoft OCR, and ignores any OCR settings in the portfolio's document types.

### Notes

You can use the Query Group method's [Multimodal Engine parameter](/senseml-reference/llm-based-methods/query-group#parameters) as an alternative to OCR engines to extract from non-text images or from poor-quality text images, such as handwriting.