What's new in 1.8

• Desktop & window screenshots — capture another app's window, another tab, or a whole screen, not just the page you're on. You pick what to capture in Firefox's own sharing prompt, so no extra permission is needed, and the shot opens in the normal editor.
• Record video & GIFs — record a tab, a window or your whole screen (you pick it in Firefox's own sharing prompt), with your microphone if you like. Pause and stop from a small floating bar you can drag anywhere. Trim the clip and save it as a WebM video or an animated GIF, or send any frame to the editor. Recorded and converted on your device.
• Smart crop — one click fits the crop box to your content, trimming blank or solid-color margins. Press Enter to accept, or adjust the handles first. Esc cancels.
• Object blur brush — paint over an avatar, logo, name, or email address and the blur follows that object's outline instead of a rectangle. A partial stroke is enough. With Smart Redact on, faces get a face-shaped blur.
• Cloud AI, your own key (off by default) — Firefox has no built-in AI model, so CaptureKit could never offer captions or rewrites here. Now it can, if you want it to: add an API key you already have from Anthropic, OpenAI or Google Gemini and you get Caption (a one-line description for alt text or a filename), ✨ Improve (rewrites your annotation text) and Ask… (questions about the capture). This is the one feature that sends a capture off your computer, so it is off until you switch it on and paste a key, Firefox asks permission before the first request, and the editor names the provider before the first upload. Your key is stored on that computer only. Smart Redact never uses it.
• Sync your settings (off by default) — carry your preferences, style presets and shape templates to every Firefox you sign into, using your Firefox Account’s own sync. No CaptureKit account and no server of ours. Your API keys, captures and history never sync, and your Smart Redact rules only if you switch that on separately.
• 19 more OCR languages (off by default) — Smart Redact and Auto-number can now read German, French, Spanish, Russian, Arabic, Hindi, Chinese, Japanese, Korean and more. Turn on Settings → AI features → More OCR languages and click Download next to the languages you need (about 1–3 MB each). Nothing is downloaded until you click, and each file is verified before it's kept.

What's new in 1.7

• Smart Redact (on-device, off by default) — one click finds emails, phone numbers, card numbers, IP addresses, IBANs, and faces in your screenshot and proposes a redaction over each one. Accept or skip each proposal, or accept them all in one step. Nothing is hidden until you accept, and nothing leaves your computer: the text reader and face detector are built into the add-on. Turn it on in Settings → AI features.
• Auto-number — with the step tool, choose "All buttons" or "All links" and numbered markers are proposed over the page's buttons or links in reading order.
• SVG export — download as a vector SVG. Text, arrows, shapes, and callouts stay editable in Inkscape or Illustrator.
• Shape templates — save a watermark, badge, or set of steps once and drop it onto any capture. It lands in the same corner, even on a different-size screenshot.
• Faster editing — keys 1–9 switch tools, hold Space to pan around long captures, and shapes snap to each other's centers and the screenshot's edges.
• Text upgrades — superscript and subscript (m², H₂O, footnote¹), taller background pills with top/middle/bottom alignment, and double-click to edit text you've already placed.
• Recolor what's selected — picking a color now recolors every selected shape, and a 🎨 button appears right next to the selection.
• The eyedropper now picks up a shape's exact color, not the blend with the screenshot underneath.

Fixes: the Settings page no longer shows a misplaced Drive note, and small layout and wording fixes throughout.