The OCR-Searchable Library

How indexing works

As soon as a capture lands on disk — from a screenshot, a recording thumbnail, or a file you drag into the Library — ScreenClipr runs a background text-recognition pass over it using Apple's Vision framework, then writes the recognized text into a local full-text search index (SQLite FTS5) alongside the capture's filename, tags, and source app. A launch-time sweep also catches any capture on disk that somehow missed indexing, so nothing is permanently skipped.

This all happens locally. Nothing about a capture or a search is ever sent anywhere — see the Privacy Policy for the full data handling picture.

What gets indexed

• OCR'd text — every word Vision can recognize inside the image itself, not just visible captions or filenames.
• Filename — including any name you've given it via Rename.
• Tags — any tag you've manually assigned.
• Source app — the application ScreenClipr detected the capture was taken from.

If OCR genuinely can't read anything in an image (a blank frame, a corrupted file), that capture is still indexed by filename — it just won't turn up for a text search that has nothing to match against.

Searching your library

1. Open the Library and focus search

   Press ⌘F to jump straight into the search field from anywhere in the Library window.

2. Type any word you remember

   It doesn't need to be the filename — search matches text that appears inside the image. Searching for a word like "oxygen" finds a screenshot with that word visible in it, even if the file is still named Screenshot 2026-09-19.png.

3. Scan the match snippets

   Results show an OCR match snippet under each card while you're searching, so you can see exactly which bit of text matched before you open anything.

How results are ranked

A search doesn't just return everything that matches — it ranks results in three tiers, so a deliberate signal always outranks incidental text:

So if you've tagged something "invoice" and it also happens to contain the word "invoice" in its OCR'd text, the tag doesn't get drowned out by unrelated captures that merely mention the same word somewhere in the image.

Filtering beyond search

The Library sidebar lets you narrow the grid without typing anything:

• Screenshots, Recordings, or Favorites as top-level filters.
• Tags you've assigned.
• Source app — captures are organized automatically by the app they were taken from.

Working with what you find

• Return opens the selected capture in the editor; Space gives you a Finder-style Quick Look without leaving the grid.
• Reveal in Finder works on a single capture or an entire multi-selection.
• Export… on the multi-select action bar copies captures anywhere on disk with collision-safe names.
• Rename… is an anchored popover that renames in place — the extension, OCR text, tags, favorite status, and date all survive the rename.
• Delete/Backspace moves the selection to Trash; the confirmation toast carries an Undo action that restores the file and reindexes it.
• Drag image or video files into the Library window to import them — they're copied in and OCR-indexed exactly like a native capture.