OCR
Turn any on-screen text into selectable, copyable content with a single shortcut. Powered by on-device macOS Vision.
How to Use OCR
- Press ⌘⇧2 to activate OCR capture mode.
- Select the screen region containing the text you want to extract.
- Snapzy instantly recognizes the text and copies it to your clipboard.
- Paste the text anywhere with ⌘V.
Pro tip — OCR works on images, videos, PDFs, and even text inside 3D apps or games. If you can see it on screen, Snapzy can read it.
Supported Scripts & Languages
Snapzy leverages the Apple Vision framework, which supports a broad set of languages and writing systems:
- Latin-based languages (English, Spanish, French, German, Vietnamese, etc.)
- CJK scripts (Simplified Chinese, Traditional Chinese, Japanese, Korean)
- Cyrillic (Russian)
- Arabic and Hebrew
- Devanagari and other major Indic scripts
Recognition accuracy is best when the text is clearly visible and the language is enabled in your macOS system languages.
OCR Benchmark
Benchmarked on Apr 19, 2026 with 12 samples per language using a synthetic UI/article-text corpus.
| Language | Character Accuracy | Exact Match |
|---|---|---|
| English | 100.0% | 100.0% |
| Vietnamese | 100.0% | 100.0% |
| Simplified Chinese | 99.3% | 75.0% |
| Traditional Chinese | 99.0% | 66.7% |
| Spanish | 99.9% | 91.7% |
| Japanese | 99.4% | 66.7% |
| Korean | 99.7% | 83.3% |
| Russian | 100.0% | 100.0% |
| French | 99.3% | 33.3% |
| German | 99.8% | 75.0% |
Real-world caveat — screenshots with emoji, low-contrast text, gradients, blur, or decorative fonts may score lower than the synthetic benchmark above.
On-Device & Private
All OCR processing happens locally on your Mac using the Vision framework. Your screen content is never sent to a server for text recognition.
- No internet connection required for OCR.
- No cloud API calls or third-party services involved.
- Works offline, even on airplane mode.
- QR detection runs alongside OCR where possible. Snapzy copies decoded payloads to the clipboard as plain text and never opens URLs automatically.
OCR in the Annotation Editor
After taking a regular screenshot, you can also right-click any image in the annotation editor and choose Extract Text to run OCR on a previously captured image.
Note — OCR capture mode does not create a file. It copies recognized text directly to your clipboard so you can paste it anywhere.