Docs

RunMacro Documentation

Create your first robot, then go further as needed.

Clear step-by-step guides to create, run, and improve robots for repetitive computer work.

Ctrl K

🔤 Image

Read on-screen text with OCR in RunMacro

Applies to: RunMacro 3.x • Updated September 2026

OCR Capture reads text in a screen region or a Browser Mode screenshot with Windows OCR. It can find target text, act at its position, and branch on the result; accuracy depends on image quality and OCR availability.

Prerequisites#

Windows OCR/winsdk is available on the runner machine.
The scan region surrounds the target text with sufficient contrast.
For Browser Mode, Chrome/CDP is connected and the target to capture is active.

Supported modes#

Desktop Mode — Captures a desktop screen region by coordinates.
Browser Mode — Uses a screenshot from the selected CDP target.

What it does#

The OCR Capture command uses Windows OCR (built into Windows 10/11) to read text from a screen region. You can enter target text to match only when that text is found, then move/click its position and branch IF/ELSE. Works in both Mouse mode and Browser Mode (CDP).

Screenshots & video#

RunMacro OCR Capture command configuration dialog showing parameter fields — Configuration dialog and parameter fields for the OCR Capture command in RunMacro.

When to use it#

Read a balance, price, or OTP shown as image/uncopyable text.
Find a label/button by its text and click it.
Check whether the screen contains a certain text to branch.

When not to use it#

The text is directly available through Smart HTML or CDP Copy Text.
The image is blurred, text is too small, or contrast is poor.
You require exact sensitive data without a separate verification step.

Configuration steps#

Add the Image › OCR Capture command.
Draw the scan region on screen (tighter around the text is more accurate).
(Optional) enter the target text to find.
Choose the found action (none / move / click / double-click / right-click) and set IF/ELSE Labels. Save.

Every option explained#

`Scan region`#

The screen area OCR reads. Draw tightly around the text for better accuracy.

`Target text`#

The string to find in the OCR result. Leave empty to take the first recognized text.

`Found action`#

None (check only) / Move / Click / Double-click / Right-click at the text position.

`IF Label / ELSE Label`#

Jump when the text is found (IF) / not found (ELSE).

`Max retries & retry delay (ms)`#

Repeat the scan to wait for the text (default 10 tries, 200ms each).

Expected result#

RunMacro locates the target text in the scan region, performs the configured action at its position (move/click/...), and follows the IF (found) or ELSE (not found) branch. It does not store the recognized text into a variable.

Real example#

Example: click a button by its text#

OCR Capture on the region with the button, target text = "Confirm".
Found action = Click.
Set an ELSE Label to branch when the text is not found.

Limitations and failure cases#

OCR can confuse similar characters and does not guarantee exact recognition.
When the OCR provider is unavailable, recognition cannot run and the command follows the ELSE branch.
DPI, capture-region, or CDP-target changes can alter the result.

Tips & common mistakes#

OCR requires Windows OCR (winsdk) to be available; if missing, the command is disabled and follows the ELSE branch. Clear, high-contrast text reads more accurately.

Verification details#

commands/capture.py
core/ocr_engine.py

← PreviousMulti Color Next →Scan OCR