Lottie vs GIF: Choose by Evidence, Not Format Hype

Start with the content, not a size multiplier

A vector Lottie icon may be much smaller than a raster frame sequence, but a Lottie file with embedded images and dense keyframes may not be. A short, optimized GIF can also be acceptable when native playback and simple distribution matter more than interaction.

Export both candidates from the same visual source and compare transferred bytes, decoded memory, first meaningful frame, CPU or frame behavior, and visual quality in the real container. Generic claims such as “five times smaller” are not substitutes for measuring the actual assets.

Capability comparison

Lottie exposes timeline controls: play, pause, direction, speed, frame seeking, markers, and segments. That makes it useful for stateful interface motion. GIF plays a frame sequence and usually needs no additional player library.

• Choose Lottie for controllable vector motion when target players support the exported features.
• Choose GIF when broad native display and a simple, non-interactive loop are the main requirements.
• Choose video for photographic or long frame-based motion where modern codecs fit better.

Compatibility and operational cost

GIF support is mature, but it offers limited playback control and can be inefficient for color-rich motion. Lottie requires a player and feature compatibility varies across web, iOS, and Android. The application must also manage player version, initialization, cleanup, and failure states.

Use the viewer's compatibility panel to identify features that deserve platform review. Then test the chosen player versions on real targets; a browser render does not settle a native-platform decision.

Accessibility and a decision record

For either format, decide whether motion is informative or decorative. Respect reduced-motion preferences, avoid harmful flashing, provide text or static alternatives when the animation communicates state, and prevent endless motion from blocking task completion.

Record why the format was selected, the measured artifact sizes, player versions, devices tested, fallback, and owner. That evidence is more reusable than a format-wide rule.