macOS Productivity Recall OCR Screenshots

OCR on Mac: How to Extract and Search Text from Images and Screenshots

A complete guide to using OCR on Mac — from Apple's built-in Live Text to building a fully searchable screenshot library with automatic text indexing.

Optical character recognition — OCR — is the technology that reads text out of images. A photo of a receipt, a screenshot of an error message, a scan of a contract: OCR turns the pixels into text your computer can index, search, and copy.

Mac has had remarkably good OCR capability since macOS Monterey, but most users only use a fraction of it. This guide covers everything: what's built in, where it falls short, and how to build a screenshot library where every image is fully searchable by its content.

What OCR Is Actually Used For

Before getting into how, it's worth being specific about why. The most common real-world use cases:

Screenshot search. You screenshot an error message, a Slack conversation, a pricing table, a design mockup. Weeks later you need to find it. Every standard search tool — Spotlight, Finder — is blind to the text inside the image. OCR-indexed screenshots change this.

Receipt capture. A photo of a paper receipt is useless in a file manager. OCR turns it into a searchable document: find "Costco" or "April" or the amount and the receipt appears.

Code and terminal output. Developers screenshot terminal output, stack traces, and API responses constantly. OCR means you can search for a specific error code across hundreds of screenshots rather than scrolling through thumbnails.

Scanned documents. PDFs and images from scanners are often image-only — the text looks right but isn't selectable. OCR extracts the text layer, making the document searchable and copy-paste friendly.

Web pages and articles. Screenshots of web content before a page goes down or a paywall goes up. With OCR, you can search for specific phrases you remember from the content.

What's Built Into macOS: Live Text

Apple introduced Live Text in macOS Monterey and has expanded it significantly since. In the Photos app, Preview, and Quick Look, you can select and copy text directly from images — Live Text runs OCR in real time and overlays an interactive text layer.

What Live Text does well:

  • Instant, real-time text recognition in Photos and Preview
  • Copy/paste from any image in those apps
  • Works with handwriting, not just printed text
  • Translates recognised text directly
  • Runs entirely on-device using Apple's Vision framework

Where Live Text falls short:

Live Text is interactive but not searchable. You can open an image and copy the text out of it, but you can't search "TypeError" across your entire screenshot library and get results. Each image has to be opened individually. There's no index.

Additionally, Live Text only works in apps that have been updated to use the API. Finder's standard search, Spotlight's default behaviour, and most third-party file managers don't benefit from it.

For occasional use — "I need to copy this address out of this photo" — Live Text is excellent. For finding things across a large screenshot collection, it's not the right tool.

The Clipboard and Drag-to-Text Approaches

Two other quick options before getting into proper indexing:

Screenshot → copy text: In Preview, open a screenshot and use the text selection tool. You can drag to select recognised text and copy it. Useful for a single extraction; not scalable.

Automator and Shortcuts: macOS Shortcuts can run OCR on images via the "Extract Text from Image" action. You can build a workflow that takes a folder of screenshots and outputs a text file of recognised content. This is scriptable and free, but requires setup and doesn't create a searchable index.

Third-party utilities: Several apps on the Mac App Store offer OCR functionality as a clipboard tool — you trigger a hotkey, draw a selection box on screen, and the recognised text is copied to the clipboard. Good for one-off extractions.

None of these approaches create an indexed, searchable library. For that you need something that watches a folder and continuously indexes new files.

Building a Searchable Screenshot Library

The practical solution for most Mac users is an app that:

  1. Watches a folder (usually Desktop and/or Screenshots) for new images
  2. Runs OCR on each new file automatically
  3. Stores the extracted text in a local full-text search index
  4. Lets you search across all indexed images from a single interface

This is what macOS doesn't provide out of the box but can be added with a third-party app.

When set up correctly, the workflow becomes invisible. You take a screenshot as normal. Within a few seconds it's indexed. Any future search that includes a word from that screenshot will surface it.

What to Look For

On-device processing. OCR involves sending images through a recognition engine. If that engine is a cloud service, your screenshots — which may contain passwords, private messages, financial information, or sensitive work — are leaving your machine. An on-device solution (using Apple Vision) keeps everything local.

FTS (full-text search) support. Basic search looks for exact strings. Full-text search with prefix matching lets you find "authenticate" by typing "auth". It's a meaningful quality-of-life difference when you can't remember the exact phrasing.

Automatic categorisation. Screenshots naturally fall into types — code, receipts, error messages, web pages. An app that classifies automatically means you can filter by type without manual tagging.

No subscription. A screenshot indexer is always-on background utility. Paying monthly for that indefinitely is hard to justify; a one-time purchase is the right pricing model.

Practical OCR Workflow Tips

Take better screenshots. Crop tightly to the content you want to find later. A screenshot of a full browser window is harder to search accurately than a screenshot of the specific table or error message you care about. ⌘⇧4 gives you a selection-based screenshot; ⌘⇧4 then Space gives you a window capture.

Name screenshots meaningfully when you care. Auto-naming by date is fine for volume, but for screenshots you know you'll need again, rename them immediately. The filename is searchable in Spotlight even when the content isn't.

Don't let the library grow without limits. Screenshot libraries accumulate fast. Periodically archive or delete screenshots you no longer need. A smaller index searches faster and surfaces more relevant results.

Use tags if your tool supports them. Automatic categorisation handles broad types. Manual tags handle specific projects or contexts: "client-X-feedback", "bug-report-sprint-12", "tax-2025". A few seconds of tagging when the context is fresh saves minutes of searching later.

Privacy Considerations

OCR on screenshots is worth thinking about from a privacy perspective. Your screenshots almost certainly contain things you wouldn't want accessible beyond your device: passwords in terminal output, sensitive messages, financial documents, client data, personal photos.

Before using any OCR app, check:

  • Does it run entirely on-device or send images to a server?
  • What is stored and where? (A local SQLite file is auditable; a cloud database is not)
  • Is the app sandboxed? What permissions does it request?
  • Is there a clear privacy policy for an app that processes sensitive screen content?

On-device OCR using Apple Vision has the best privacy profile — Apple's framework runs locally and the app never needs a network connection to function.

Recall is a Mac app that automatically indexes your screenshots using on-device OCR — everything runs locally using Apple Vision, the index lives in a SQLite database on your own machine, and nothing is sent to any server. Search your entire screenshot library by content in seconds.

Download on the Mac App Store

Free on the Mac App Store.

Back to Blog