Tags · greensock/GSAP

- NEW: Thanks to Webflow, GSAP is now **100% FREE** including ALL of the bonus plugins like SplitText, MorphSVG, and all the others that were exclusively available to Club GSAP members. That's right - the entire GSAP toolset is FREE, **even for commercial use!** 🤯 So now the public repository has all of the bonus plugins. You can read more about this on Webflow's blog: https://webflow.com/blog/gsap-becomes-free

- NEW: SplitText was completely rewritten, 50% smaller and 14 new features. There were a few VERY uncommon breaking changes. See https://gsap.com/blog/3-13 for details.

- NEW: you can animate a CSS value to something like "var(--my-variable)".

- IMPROVED: if you set the "ignore" property of an Observer, it now ignores descendants of those elements too.

- IMPROVED: modifiers plugin and endArray plugin will work even in headless environments (where there's no "window" defined). See https://gsap.com/community/forums/topic/44301-nodejs-gsap-modifiers/

- IMPROVED: GSDevTools has the new GSAP branding.

- FIXED: if you use Pixi.js version 8+ and attempted to animate more than one filter on an element, PixiPlugin could throw an error.

- FIXED: if you try to Flip an element that's in the shadowRoot (like in a Web Component), it may not calculate the position correctly. See https://gsap.com/community/forums/topic/44135-gsap-web-components/

- FIXED: drawSVG may appear to jump near the very end or start if the path is adequately short or uses very small numbers. See https://gsap.com/community/forums/topic/44072-drawsvg-on-mask-issue/

- FIXED: if you Flip an element that has transforms applied only via a CSS rule (not via GSAP or inline), it may not interpret it correctly. See https://gsap.com/community/forums/topic/44321-bug-on-fixed-position-using-flip/

- FIXED: if you change the timeScale of a timeline that has a cached duration value, it may incorrectly calculate the time. See https://gsap.com/community/forums/topic/44232-if-globaltimelinetimescale-is-set-immediately-after-timeline-played-the-start-of-the-animation-will-be-delayed/

- FIXED: typo in the gsap.utils.mapRange() method. See #615

- FIXED: a Flip.fit() with a duration set may immediately apply the extra props. See https://gsap.com/community/forums/topic/44398-flipfit-props-not-transitioning/

- FIXED: regression in 3.12.6 that could cause a ScrollTrigger on an animation with immediateRender: false not to initialize with the correct value. See https://gsap.com/community/forums/topic/44387-initial-states-of-fromto-tweens-in-a-scrolltrigger/

- FIXED: if the playhead of a repeated animation traveled from the very start all the way past the first iteration, its onStart wouldn't be called. See https://gsap.com/community/forums/topic/44466-onstart-not-called-when-on-low-fps/

- FIXED: from() and fromTo() tweens inside a timeline that hasn't been rendered yet but gets reverted wouldn't have their initial (immediateRender) values reverted. This manifested in ScrollTriggered timelines that had from() tweens and invalidateOnRefresh: true not being rendered correctly. See https://gsap.com/community/forums/topic/44365-how-to-properly-clear-cached-tweens-after-resize/

Apr 30, 2025
9f8eca1
zip
tar.gz

3.12.7

3.12.7

- FIXED: regression in 3.12.6 that set pointer-events: none on ScrollTrigger pin spacers which caused descendents to stop responding to clicks and other pointer events.

- FIXED: regression in 3.12.6 that could throw an error if you don’t pass a vars/config object to Flip.fit().

Jan 15, 2025
5e13097
zip
tar.gz

3.12.6

3.12.6

- IMPROVED: if you set immediateRender: true on a motionPath tween, it'll force the initial render to put that element on the path right away.

- IMPROVED: ScrollTrigger.sort() uses a better default algorithm that prioritizes the vertical position of the trigger element instead of the calculated start position of the ScrollTrigger. See https://gsap.com/community/forums/topic/40584-scrolltriggerrefresh-doesnt-do-anything-after-pinned-sections-are-dynamically-loaded/

- IMPROVED: TypeScript definition for MotionPathPlugin, like for allowing strings in the Array-based syntax, like [{x: "100vw"}] and function-based values for start/end. See https://gsap.com/community/forums/topic/40894-typescript-errors-for-start-and-end-fields-for-motionpath/ and #578

- IMPROVED: you can pass true to ScrollTrigger.refresh(true) to have it do a "safe" refresh, meaning that if the page is in the middle of scrolling, it'll wait until it's done before doing the refresh. That way, it won't kill an in-progress momentum scroll. See https://gsap.com/community/forums/topic/40935-gsap-scrolltrigger-start-position-breaks-on-dom-changes/

- IMPROVED: added pointer-events: none to pin spacer elements that ScrollTrigger creates which helps work around a Chrome browser bug.

- IMPROVED: Expo ease was enhanced to eliminate the possible jump at the very end. See https://gsap.com/community/forums/topic/25331-weird-jump-at-end-of-tween-with-expoout-ease/

- FIXED: if you set a zIndex on a Flip.fit(), it works now.

- FIXED: reading transform-related properties on a position: fixed element that has no transforms applied could result in the element getting temporarily added to the DOM at the root which could cause iframe content to reload, like a Vimeo video. See https://gsap.com/community/forums/topic/39645-flip-with-iframe-causes-iframe-to-rerender/

- FIXED: repeatRefresh may not actually refresh in a very rare condition (regression in 3.12.3).

- FIXED: if you use special characters like >,<, &, etc. in a TextPlugin tween on an SVG element, those characters could get replaced with the HTML code (like &gt; for >). See https://gsap.com/community/forums/topic/35550-handwritng-and-typewriter-animation-on-path-in-svg/

- FIXED: if you gsap.registerPlugin(ScrollTrigger) in an SSR framework like Next.js, it could complain that a style attribute was added to the document.body. That's removed now.

- FIXED: if you set a non-zero stagger value in gsap.defaults() (which is a very strange thing to do), it could throw an error when creating a gsap.quickTo().

- FIXED: repeatRefresh is in the TypeScript definitions now for staggers. See #595

- FIXED: on a motionPath tween, if you use an Array-based path with targets that have a "style" property but are NOT DOM nodes, and the properties have names that are transform-related like x, y, scaleX, scaleY, etc., it could throw errors. See https://gsap.com/community/forums/topic/40023-motionpath-plugin-problem/

- FIXED: if you navigate back to a page that then reuses the same ScrollTrigger instance (like a React routing change), a timeline that has a scrubbed ScrollTrigger may not render the playhead at the right in-progress spot. See https://gsap.com/community/forums/topic/39982-scrolltrigger-breaks-when-navigating-back-usegsap-react-react-router/

- FIXED: on an Observer, if you set passive: false that may be ignored unless preventDefault is true. This is fixed now.

- FIXED: ScrollTriggers with a containerAnimation may not refresh their start/end values correctly if ScrollSmoother is used or ScrollTrigger.sort() is called. See https://gsap.com/community/forums/topic/37419-horizontal-scroll-resize/

- FIXED: on a tween with repeatRefresh: true, it may skip the refreshing of values if the parent playhead lands beyond where the repeat spot is, but by less than 0.000001 seconds (exceedingly rare!).

- FIXED: if you do a .from() animation of borderRadius using a unit of "rem", it may jump a bit at the very end of the tween. See https://gsap.com/community/forums/topic/40096-scroll-trigger-auto-scroll-back-and-animation-over-shoot/

- FIXED: if you revert() a gsap.context() that has a ScrollTrigger with a scroller that's not the window, it could cause snapping to no longer function. This affects the @gsap/react package's useGSAP() hook too. See https://gsap.com/community/forums/topic/40198-snapping-breaks-when-re-rendering-in-react/

- FIXED: if you set once: true on the first ScrollTrigger (in terms of refreshPriority order) and it activates on page load (meaning its start is smaller than the page's current scroll position), it could throw an error. See https://gsap.com/community/forums/topic/40242-scrolltriggerbatch-bug-with-once-scrolltriggers/

- FIXED: if you set a minimumMovement on a Draggable of type: "rotation", it would interpret it in degrees rather than pixels, and a minimumMovement of 0 would actually default to 2 (degrees).

- FIXED: worked around an odd behavior in Firefox where the browser would fire pointermove events without the pointer moving at all which affected Observer's lockAxis behavior and it could prematurely fire an onDrag, etc.

- FIXED: if you kill() a completed tween that has a ScrollTrigger, it may not also kill the ScrollTrigger. See https://gsap.com/community/forums/topic/40359-gsap-kill-doesnt-stop-scrolltrigger-onupdate-from-running/

- FIXED: extremely rare edge case where a lazy-initted tween is initially set to render at a non-zero time but before that lazy tween renders lazily, a render is triggered at a time of 0, it could end up rendering at the wrong time (the lazy-initted one instead of the slightly later render).

- FIXED: orientation changes didn't properly trigger a ScrollTrigger.refresh(), so the start/end values of ScrollTriggers could be inaccurate after an orientation change. See https://gsap.com/community/forums/topic/40713-scrolltriggers-are-positioned-wrong-after-orientation-change/

- FIXED: an Observer may fire an onDrag() AFTER the onDragEnd() if debounce is true, but now it'll make sure it fires it BEFORE the onDragEnd(). Also, with debounce: false, an onDrag() could fire before the onDragStart(). See https://gsap.com/community/forums/topic/40909-observer-ondrag-firing-after-ondragend/

- FIXED: in a very uncommon scenario, creating ScrollTriggers inside a DOMContentLoaded event might cause the start/end values to be miscalculated. See https://gsap.com/community/forums/topic/41267-scrolltrigger-breaks-on-refresh-when-using-domcontentloaded/

- FIXED: MotionPathHelper.editPath() may throw an error if you add an anchor to a path that has MULTIPLE segments (more than one "M" command).

- FIXED: if you apply a snap to a ScrollTrigger that has a non-scrubbing animation, and the start is at 0 or less, it might restart the animation after the snap duration elapses on page load.

- FIXED: if you remove() an animation from a timeline that was NOT its parent, it could cause an error.

- FIXED: if you call CustomEase.get() in an SSR environment before the window object is defined, it could throw an error. See https://gsap.com/community/forums/topic/42510-gsapregisterplugin-in-nodejs-server/

- FIXED: some links in the TypeScript definitions file were incorrect since the gsap.com site launch.

- FIXED: in very rare cases, a CustomEase might end at something like 0.999999999 instead of 1 progress-wise (so values wouldn't quite land at the destination). That's resolved now.

- FIXED: if you clearProps on a target that contains individual transform component values applied in a CSS rule (like translate, rotate, or scale), it may not properly clear the transform. See https://gsap.com/community/forums/topic/43073-separate-transform-properties-like-rotate-and-scale-behave-differently-then-a-combined-transform/

- FIXED: regression in Flip.fit() could cause the width/height of the initial state not to be factored in, but only if there was a non-zero duration.

- FIXED: an Angular issue related to SVG elements that return an empty getBBox() which could cause an error due to the temporary reparenting and re-assigning the getBBox() method itself inside GSAP. See https://gsap.com/community/forums/topic/42859-gsap-321-override-of-the-svg-api-method-getbbox-removes-angular-dblclick-binding-from-dom-element/

- FIXED: if you disable() and then change the scroll position and then enable() an Observer, on the next scroll it could factor in the pre-disabled scroll position in the calculations, thus firing an onDown() or onUp() (for example) incorrectly.

- FIXED: if you restart() a zero-duration timeline that has an onComplete which already fired, it will fire that again after the restart(). See #604

- FIXED: zero-duration tweens wouldn't resolve then() Promises. See #590

- FIXED: TypeScript definition for getById() allows for undefined as a return type. See #589

- FIXED: ScrollSmoother doesn't set overscroll-behavior: none on the HTML/BODY anymore due to a Safari bug that caused that to stop the page from scrolling at all. See https://gsap.com/community/forums/topic/43802-loading-maps-api-scrollsmoother-in-safari-173-breaks-mousewheel-scroll-for-the-page/

- FIXED: PixiPlugin better accommodates the changes in Pixi.js version 7+, like fillColor and strokeColor. See #580

Jan 15, 2025
f4f16d6
zip
tar.gz

3.12.5

3.12.5

- IMPROVED: ScrollTrigger's ignoreMobileResize is true by default now because a few versions back we switched to calculating the window height based on 100vh instead of window.innerHeight, so it really isn't necessary to force a .refresh() on mobile resize (address bar showing/hiding). This change improves performance on mobile devices because it skips the expensive refresh() when the address bar shows/hides.

- IMPROVED: removed a DisplayObject check in PixiPlugin so that it's more compatible with the upcoming release of Pixi.js version 8. See #561

- IMPROVED: the anticipatePin feature works when scrolling in either direction (previously it worked when scrolling forward only). See https://gsap.com/community/forums/topic/39521-anticipatepin-in-reverse/

- FIXED: regression in 3.12.3 could cause a ScrollTrigger that has a numeric scrub value and a snap to prematurely jump to the final progress value when the snap finished rather than waiting for the scrub to finish. See https://gsap.com/community/forums/topic/39363-scrolltrigger-scrub-clip-path-flicker-when-scroll-back/ and #567

- FIXED: regression in 3.12.3 could cause a tween with repeatRefresh: true AND a non-zero repeatDelay not to refresh. See https://codepen.io/GreenSock/pen/yLwLgNQ/db88d564d6308b9fcff7b65efb95febf?editors=1010

- FIXED: if you set a ScrollSmoother's effects() in a delayed fashion (after two ticks elapse after page load), it may not calculate offsets properly. See https://gsap.com/community/forums/topic/39380-scroll-smoother-effects-targets-jump-on-effectstrue/

- FIXED: regression in 3.12.3: if the playhead lands exactly on top of the yoyo position of a keyframed tween, it may render at the state before the animation started (just for 1 tick). See #558

- FIXED: if you set invalidateOnRefresh: true on a ScrollTrigger that's attached to a .fromTo() tween, it may not render at the initial state on refresh(). See https://gsap.com/community/forums/topic/39386-svgdraw-window-resize-issues/

- FIXED: if you create a gsap.matchMedia() inside a gsap.context() and then revert() that Context, it may not clear out non-matching media queries (they may fire again when matching). See https://gsap.com/community/forums/topic/39481-pin-spacer-is-double-the-height-it-should-be-after-window-resize/

- FIXED: if the system clock changes backwards to an earlier time, it could cause animations to stop. See #568

- FIXED: if a flexbox child is pinned by ScrollTrigger and pinSpacing is false, flex-basis is now set to the pixel height/width of the element to avoid common layout problems.

- FIXED: missing TypeScript parameter for ScrollTrigger.clearScrollMemory(). See #571

- FIXED: if you set a default stagger value via gsap.defaults(), it could cause an error in ScrollTrigger if you create one with a numeric scrub. See https://gsap.com/community/forums/topic/39600-scrolltrigger-numeric-scrub-throws-console-errors-in-gsap-core-cannot-read-properties-of-undefined-reading-totalprogress/

- FIXED: if a motionPath tween had a start or end value that wraps around and would land precisely on an anchor point that is not at the very start or end of a path, it could render incorrectly. See https://gsap.com/community/forums/topic/39594-why-are-content-1-and-content-2-overlaped-also-what-is-the-best-way-to-add-a-timeline-to-a-main-tween-or-timeline/

Jan 18, 2024
f836a3f
zip
tar.gz

3.12.4

3.12.4

- FIXED: regression in 3.12.3 could cause ScrollTriggers in gsap.matchMedia() to stop functioning after quick resizes that pass the matchMedia bounds. See https://gsap.com/community/forums/topic/39241-animations-freeze-after-a-few-window-resizes-and-using-gsapmatchmedia/

- FIXED: overwriting a scrollTo tween in such a way that didn't leave any more properties tweening, onInterrupt() wasn't invoked.

- FIXED: SplitText TypeScript definitions file needed an "elements" property. See #564

- FIXED: if you Promise.all() or .then() a timeline that already finished and has a zero duration, that Promise may not resolve properly. See https://gsap.com/community/forums/topic/39291-promiseall-dont-resolve-when-overwrite/

- FIXED: regression in 3.12.3 could cause transforms (x, y, scaleX, scaleY, rotation, etc.) on SVG elements to calculate the origin incorrectly. See https://gsap.com/community/forums/topic/39197-update-3123-broke-my-code-for-animating-clip-paths and #563

Dec 14, 2023
3d86e43
zip
tar.gz

3.12.3

3.12.3

- IMPROVED: gsap.context() and gsap.matchMedia() functions will get a 2nd argument that we'll call "contextSafe" which is like a wrapper for any function that you'd like to keep in the context. So any GSAP animations/ScrollTriggers/Draggables/Observers that are created during the execution of that function will be added to the context and the selector text will be scoped.

- IMPROVED: added a suppressEvents parameter to Tween/Timeline/Animation .timeScale() method so that you can optionally prevent it from suppressing events when altering the timeScale. See #551

- IMPROVED added an ignoreSpeed [3rd] parameter to ScrollSmoother's offset() method to allow you to specify whether you want to get the value that corresponds to the window's scroll position or the ScrollSmoother's scrollTop value. See https://greensock.com/forums/topic/35108-problems-with-the-scrollsmoother-plug-in-speed-option-in-gsap-3114/

- IMPROVED: Observer now fires an onStop with an onMove (previously it would only fire after a press). See https://greensock.com/forums/topic/38469-observer-misunderstanding-with-onchange-onmove-and-onstop/

- IMPROVED: slight change to TypeScript definitions for EaseString allows arbitrary strings while also activating code hinting for the common ones in more environments. See #556

- IMPROVED: animations/ScrollTriggers created inside a ScrollTrigger's callback like onEnter/onLeave/onToggle/onEnterBack/onLeaveBack will be added to the original Context (if one existed), meaning selector text will be scoped properly. See https://gsap.com/community/forums/topic/38850-reactgsap-why-i-cant-use-class-selector-in-scrolltrigger-inside-gsapcontext/

- IMPROVED: a ScrollTrigger's snap end position is limited to the resolution of the browser's scroll (whole pixels only), thus if you have a scrubbed animation that's supposed to snap to a very specific spot on that animation, it may end slightly off of that but now a correction runs at the end of the snap to ensure that it gets set PRECISELY to that snapped position. See https://gsap.com/community/forums/topic/38937-scrolltrigger-timeline-snapping-with-label-doesnt-snap-precisely-to-label-position/

- IMPROVED: the gsap-trial files will now work on domains that end in ".local" (for testing only please)

- FIXED: for an Observer, if you pressed, started dragging but released within 3 pixels of the original press, it wouldn't fire the onDragEnd. See https://greensock.com/forums/topic/37510-the-problem-that-ondragend-is-not-called-in-observer/

- FIXED: if you revert() a context/matchMedia that has a reversed animation, it may not render things in the proper order, potentially leaving inline styles when it shouldn't. See https://greensock.com/forums/topic/37432-issues-with-horizontalloop-helper/

- FIXED: if Draggable is applied to an <object> where the document isn't defined initially, an error could be thrown. See #549

- FIXED: if you apply a speed to a ScrollSmoother (other than 1), the offset() method would return a value that corresponds to the window's scroll position rather than the ScrollSmoother's scrollTop (which is affected by speed). See https://greensock.com/forums/topic/35108-problems-with-the-scrollsmoother-plug-in-speed-option-in-gsap-3114/

- FIXED: if you apply an onUpdate to a .fromTo() tween, the scope ("this") would be linked to the zero-duration "from" tween instead of the actual tween instance you'd expect, thus its progress() would be 1 instead of 0 at the very start.

- FIXED: if you define position: absolute on a SplitText and then revert(), the width/height inline styles could remain instead of getting cleared out. See https://greensock.com/forums/topic/38391-responsive-behavior-using-splittext/

- FIXED: if you create a ScrollTrigger with a "snap" and pin inside a gsap.matchMedia(), it could lead to incorrect positioning of the pinned element after a resize that makes it no longer match.

- FIXED: if you set a CSS value to "auto" (like height or width) via a GSAP tween and then you revert() it, the original computed value could be left as an inline style instead of cleared out.

- FIXED: if you call kill() or revert() on a MotionPathHelper in certain specific conditions, it may delete the original <path> itself.

- FIXED: worked around a very rare scenario where document.createElement() returns undefined and would consequently throw an error in CSSPlugin. See #553

- FIXED: if you create a staggered animation inside a gsap.context() or gsap.matchMedia() that affects the same element(s) as a gsap.set() that happened before that staggered animation and then the context/matchMedia gets reverted, the initial value may not get reverted properly.

- FIXED: if you animate a percentage-based width/height of an element whose parent has padding and/or is flexbox/grid, and the target element doesn't have that property already set inline, it may miscalculate the starting value of the tween. Related to https://gsap.com/community/forums/topic/38599-how-to-animate-object-fit/

- FIXED: if you create multiple SplitText instances on the same element inside a gsap.context() and then revert() that Context, it may not fully revert the element to its original state. See https://gsap.com/community/forums/topic/38734-splittext-innerwrap-renders-twice-in-strictmode/

- FIXED: if you set allowNestedScroll: true in the ScrollTrigger.normalizeScroll() feature, touch-scrolling on a link on a mobile device could result in a click event firing on that link. See https://gsap.com/community/forums/topic/38770-why-does-the-burger-menu-scroll-along-with-the-content-on-mobile/#comment-193009

- FIXED: if you set end: "max" or clamp() the end of a ScrollTrigger that has a pin, and the pinSpacing was extending the page taller, the dynamic adjustment of the end value would also affect the pinSpacing, reducing the maximum scroll area which then wouldn't be accurately reflected in the final end value.

- FIXED: if you define a stagger with grid: "auto" on an Array of elements that don't wrap at all (not really a grid), the last element's timing wouldn't be correct. See https://gsap.com/community/forums/topic/38536-when-scrolling-down-a-batch-the-stagger-doesnt-seem-to-work-correctly/

- FIXED: if you clamp() a ScrollTrigger's start value and it would naturally (without clamping) resolve to beyond the maximum scroll position, it wouldn't get clamped. Starting values were only clamped such that they weren't allowed to be negative (focused on the top of the page only, not the bottom too).

- FIXED: if you dynamically added/created a ScrollTrigger while ScrollSmoother was mid-scrub, it could lead to the scroll jumping. See https://gsap.com/community/forums/topic/37515-dynamic-scrolltrigger-with-pin-inside-a-scrollsmoother/

- FIXED: if you run SplitText on text that has words separated by non-breaking spaces (&nbsp;), it wouldn't recognize that as a word delimiter. See https://gsap.com/community/forums/topic/37271-why-does-the-splitted-node-from-splittext-includes-extra-whitespace-note-only-the-node-which-is-next-tag/

- FIXED: Flip.fit() is now gsap.context()-aware so that it'll revert inline styles when the context is reverted.

- FIXED: regression in 3.11 that could cause transformOrigin to lose its "z" portion if you apply it in a .from() or .fromTo() tween or a context that gets reverted. See https://gsap.com/community/forums/topic/38958-rotatey-not-working-in-nuxt/

- FIXED: if you call gsap.registerPlugin(Observer) multiple times, it could throw an error.

Nov 30, 2023
641da46
zip
tar.gz

PreviousNext

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

3.15.0

3.14.2

3.14.1

3.14.0

3.13.0

3.12.7

3.12.6

3.12.5

3.12.4

3.12.3

Tags: greensock/GSAP