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

# OCR

> Extract text from images and selected screen regions.

Enconvo provides OCR commands and a selectable OCR provider. Use local Apple OCR, local MLX OCR, an AI model, or a supported cloud OCR provider according to the material you need to read.

## Extract text from the screen

1. Open **Settings → Global Providers Settings → OCR** and choose a provider.
2. Run **Screenshot OCR** and select the relevant screen region.
3. Review the extracted text before copying or using it in another command.

**Silent Screenshot OCR** copies its result to the clipboard. **OCR** processes supplied image input. **Screenshot Translate** and **Screenshot Explain** add a follow-up task to a screen capture.

## Choose a provider

| Provider | Processing |
| - | - |
| Apple Native OCR | Native local text recognition |
| PP-OCRv6 (MLX) | Local detection and recognition on supported Apple Silicon hardware |
| AI Model OCR | Use a configured AI model and instruction |
| Mistral OCR / Cloud variant | Remote OCR through configured access |

See [Provider Catalog](/providers/catalog#ocr) for current configuration fields.

## Improve accuracy

Capture a tight, readable region at a useful scale. Check columns, tables, punctuation, code indentation, and ambiguous characters manually. A scanned document may require OCR even when it has a PDF file extension.

If screen capture is blank or unavailable, check [Permissions](/configuration/permissions). If recognition fails, test a simple image and read the selected provider's error. For asking about a visual element rather than extracting text, try [Appshots](/features/appshots) or [Screen Doodle](/features/screen-doodle).


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.