Image Text Selector 作成者: Foivos

macOS only — requires a Mac and a small free companion helper app (one-command install).

Image Text Selector makes the text inside images selectable, exactly like normal page
text. Hover over any picture that contains text — a screenshot, a scanned page, a chart,
a photo of a sign — and after a fraction of a second an invisible text layer appears
precisely over the printed words. Drag to select, press Cmd+C, and you get the real text
with proper line breaks. It brings functionality similar to Apple's Live Text feature in
Safari to Firefox.

Everything stays on your Mac. The extension contacts no servers of its own; its only
network activity is re-requesting the image you are hovering from the same website that
is already showing it (without your cookies). Recognition runs entirely on your device
via Apple's Vision framework, through a small open-source helper program you install
once. No telemetry, no analytics, no accounts — the data-collection declaration you see
above ("no data collected") is the whole story.

Why it asks for access to all websites: it needs to read the pixels of the image you
hover — on whatever site you happen to be — and that image is the only thing it ever
touches. Processing never leaves your machine.

Features

• Native-feeling selection: one continuous highlight band per line, drawn in the same
  screen refresh as your drag — no flicker, no hairline gaps
• Sweep selection: one diagonal drag grabs whole paragraphs
• Right-click on an image still gives you the normal image menu ("Save Image As…");
  right-click on your highlighted text gives the Copy menu
• Vertical and rotated text (margin stamps, book spines) selects and copies correctly
• Copy produces clean text with real line breaks — only what you selected
• One-click toolbar toggle; when off, the extension does literally nothing
• Engineered for zero idle cost: no background processes, no timers, strictly bounded
  memory; the OCR helper runs only for the fraction of a second each recognition takes

Requirements

• macOS 14 (Sonoma) or newer — on macOS 26+ it automatically uses Apple's newest
  document-recognition engine
• Firefox 140 or newer (matches the shipped strict_min_version)
• The free open-source helper app: install it with one command from https://github.com/fffoivos/image-text-selector
  (instructions on the page). Without it, the extension shows a setup guide and does
  nothing else.

Languages: the ~30 languages Apple's Vision engine supports, including English,
French, German, Spanish, Italian, Portuguese, Chinese, Japanese, Korean, Russian,
Ukrainian, Thai and Vietnamese. Greek is not currently supported by Apple's engine
(the same limitation applies to Apple's own tools).

Not for you if: you use Windows or Linux (the extension stays inert and says so),
or you need a language Apple's engine doesn't cover.

Source code (extension and helper): https://github.com/fffoivos/image-text-selector

Apple, Safari, Live Text, macOS and Apple Vision are trademarks of Apple Inc.,
registered in the U.S. and other countries and regions. This extension is an
independent project, not affiliated with or endorsed by Apple.