Rendered by GPUI’s web backend from the real component. Requires a browser with WebGPU.
avatar.rs presents a circular image
with explicit alternative text and application-supplied fallback text:
Avatar::new("preset-avatar", "Balanced preset artwork", "BL", palette)
.image(preset_artwork)
.size(rems(2.5))image accepts GPUI's existing ImageSource conversions. Omitting it shows the
fallback immediately; an image error also shows the fallback. While an image is
pending, GPUI's native 200 ms loading delay applies before showing the fallback.
GPUI owns decoding, loading, caching, and redraws. The recipe adds no separate
loading task, timer, cache, or retry policy.
The root supplies one native Image role and the given accessible name in every
state. The image and fallback add no separate name or focus stop. Supply a name
that describes the image's meaning, rather than relying on the visible initials.
The application chooses fallback text and any localization; this recipe does not
derive initials from names. The default frame is 2.5 rem square, with palette
colors and circular clipping. Styled can change its size and colors; keep width
and height equal for a circle. Images use GPUI's cover fit. Status badges, avatar
groups, and interactive actions remain application compositions.
The example embeds preset-avatar.svg
for the Balanced preset. Other presets use explicit text such as PR and HQ,
and an empty selection uses ?, with an accessible name derived from the current
preset label. The source is selected while rendering the existing HoverCard from
the retained Combobox state; there is no second selected-preset value. This
bundled artwork requires no network request. Applications supply their own assets
when copying the recipe.
AvatarGroup::new(id, "Team members", palette) stacks .avatar(...) items like
shadcn's avatar group. Each visible avatar sits in a surface-colored ring and
overlaps the previous one. .max_visible(n) shows the first n and adds a
"+N" chip whose Label name lists the hidden people ("2 more: Cy Lee, Di Cho"),
so nobody is lost to assistive technology. The frame is a named Group with no
focus stop. The preset preview shows five people with three visible.