diff --git a/docs/ocr.md b/docs/ocr.md
new file mode 100644
index 00000000..bc1bbf09
--- /dev/null
+++ b/docs/ocr.md
@@ -0,0 +1,29 @@
+# OCR
+
+This feature allows you to select a screen region, recognize it as text and copy this text to the clipboard.
+Recognition works locally on your PiKVM and does not use any cloud services. It uses the [Tesseract OCR library](https://github.com/tesseract-ocr/tesseract).
+Tesseract does not see your image until you explicitly give the recognition command. The evil AI is not watching your screen.
+
+## Language support
+
+For any language, you will have to install its support. It is very easy to do this (Russian language for example):
+```
+# rw
+# pacman -S tesseract-data-rus
+# ro
+```
+
+List all available languages in the repository:
+```
+# pacman -Ss tesseract-data
+```
+
+## Disabling OCR
+
+If you want to get rid of this feature completely, you need to delete Tesseract and all its data:
+
+```
+# rw
+# pacman -R tesseract
+# reboot
+```