How to Extract Unselectable Text from Any Video, Image, or App on Mac
How to copy text from paused YouTube videos, error dialogs, iOS simulators, PDFs, and screenshots on macOS with instant on-screen OCR.
We have all encountered text on macOS that should be selectable, but isn't:
- A complex terminal command or shell script in a paused YouTube coding tutorial
- An unselectable error hash or stack trace inside an alert sheet or modal dialog
- A long URL or token inside a Figma mockup or PNG design spec
- Text rendered inside an iOS Simulator, remote desktop (VNC/RDP), or Docker dashboard
- A scanned PDF or locked document that disables cursor highlighting
Manually retyping 80-character UUIDs, API keys, or nested JSON structures is tedious and error-prone.
While macOS includes native Live Text inside Preview and Safari, it doesn't work inside arbitrary third-party apps, active video players, or modal dialogs. Here is how modern on-device OCR works on Mac and how to copy any text on your screen in a split second.
The Problem: Pixels Without Semantic Text Nodes
When text is rendered natively in macOS (like in TextEdit or Pages), the operating system maintains an accessible text storage object (NSTextStorage). Your cursor can drag across glyph boundaries and copy UTF-8 strings.
However, many modern workflows flatten text into raw pixels:
Sources of Unselectable Text:
┌───────────────────────────────────────────────┐
│ 1. Video Players (YouTube, Loom, Vimeo) │ ──► Pure video frames
│ 2. Remote Desktops & Simulators │ ──► Streamed raster buffer
│ 3. Figma / Canvas-based Graphics │ ──► WebGL / GPU drawn shapes
│ 4. Native Dialogs & System Crash Alerts │ ──► Non-selectable labels
└───────────────────────────────────────────────┘
│
▼
Mouse selection is disabled!
To extract this text without manually typing it out, you need a high-speed OCR (Optical Character Recognition) pipeline that hooks directly into your display server.
In-Flight OCR: The Fastest Way to Grab Text
Many users install standalone OCR utilities like TextSniper or Shottr. But running a separate utility just to read text introduces menu bar clutter and duplicate keyboard shortcuts.
The fastest ergonomic pattern is in-flight OCR mode directly inside your screenshot tool:
Screenshot Hotkey (⌘⇧1) ──► Drag box over unselectable text
│
┌──────────┴──────────┐
▼ ▼
Release Mouse Tap ⌘ Key
│ │
Save/Markup Image Extract Plain Text (20ms)
│
▼
Copied to Clipboard!
Instead of wondering "Do I need a screenshot or text?", you use the exact same capture reflex. If you want the image, you release. If you want the text, you tap ⌘.
Step-by-Step: Copying Unselectable Text with Sigght
Sigght bundles on-device Apple Vision text recognition with zero third-party cloud dependencies. Here is how to use it:
1. Trigger Capture (⌘⇧1)
Whenever you see text you need to copy—whether in a video, an Xcode simulator, or an error box—press ⌘⇧1.
The screen instantly freezes, preventing video frames from advancing or tooltips from disappearing.
2. Box the Target Text
Drag a rectangular marquee over the code snippet, email address, or error message.
3. Tap ⌘ to Grab Text
- While holding or after selecting the area, tap the ⌘ modifier (or click the Text icon on the floating HUD).
- Sigght passes the bounding box directly to Apple's Neural Engine (
VNRecognizeTextRequest) with fast recognition mode. - In less than 20ms, the extracted plain text is formatted and copied to your system clipboard.
- A subtle haptic confirmation lets you know the text is ready.
4. Paste Directly (⌘V)
Switch to your code editor, Slack, or browser and paste.
- Single-line codes, URLs, and serial numbers are cleanly trimmed without trailing linebreaks.
- Multi-line blocks retain logical paragraph and indentation structure.
1-Character Precision vs. Word Snapping
A common flaw with legacy OCR utilities is aggressive "word-boundary snapping". When you try to copy a specific terminal argument (like --dry-run or an API endpoint /v1/checkout), poorly calibrated OCR engines often snap to the whole line or include unwanted punctuation.
Sigght uses tight bounding-box anchoring:
- It evaluates character glyph confidence scores on-device.
- Leading and trailing whitespace is intelligently trimmed.
- Special developer characters—such as backticks, brackets
{}, colons:, and slashes/—are preserved without being corrupted into similar-looking letters.
Security & Privacy: Why On-Device OCR Matters
Text you capture often contains sensitive information:
- One-Time Passwords (OTPs) and 2FA recovery codes
- Internal staging URLs and authentication tokens
- Confidential customer emails and financial tables
Cloud-based OCR extensions upload full-resolution screenshots to external server clusters, exposing your workspace data to compliance risks.
Sigght executes 100% of text recognition locally on your Mac's hardware. No network requests are made, no telemetry logs are sent, and the captured buffer is immediately discarded from memory once copied.
Comparison: Screen OCR Methods on Mac
| Feature | macOS Native Live Text | Standalone Apps (TextSniper) | Sigght Toolkit |
|---|---|---|---|
| Trigger Method | Hover inside Safari/Photos | Dedicated OCR hotkey | Unified capture hotkey (⌘⇧1 + ⌘) |
| Works on Paused Video | Only in Safari native player | Yes | Yes (Any player, YouTube, Loom, Zoom) |
| Developer Syntax Accuracy | Low (Aggressive autocorrect) | Moderate | High (Preserves code syntax & symbols) |
| Menu Bar Apps Needed | Built-in | 1 Extra background app | 0 Extra (Combined with capture & video) |
| Privacy / Offline | On-device | On-device | 100% On-device (Zero network calls) |
| Price | Free (macOS 12+) | $12.00 one-time | Free Forever |
Copy Any Screen Text Instantly
Never waste time retyping unselectable text from videos, error logs, or mockups again. Sigght gives you instant on-screen OCR, auto-expanding screenshot markup, and 60fps recording in one lightweight, native Mac app.
Download Sigght for Mac — Free Forever
(v1.0.2 · macOS 14 or later · 10 MB)
Frequently Asked Questions
How do I copy text from an image or video on Mac?
Using Sigght, trigger capture with ⌘⇧1, drag a box over any text on screen, and tap ⌘ (or click the OCR toggle). Sigght uses Apple's native Vision framework to instantly extract the plain text and copy it to your clipboard.
Why is copying text from video tutorials or simulators so difficult?
Video streams, remote desktop sessions, graphics editors, and system crash popups render text purely as raster pixels rather than selectable DOM/OS text nodes, making normal mouse highlighting impossible.
Does on-screen OCR send my screen data to the cloud?
No. Sigght runs Apple's VNRecognizeTextRequest entirely on-device using the Apple Neural Engine. Your screen contents, sensitive credentials, and private documents never leave your Mac.
macOS Screen Intelligence & Sigght
Engineering guides on zero-delay screen capture, discrete zoom recording, and Apple Silicon Neural Engine OCR.