Iced Rs logo

Iced Rs

Organization
vanillagreencom
iced-rs

Load whenever building or debugging an Iced UI.

Overview

Publishervanillagreencom
Repositorykendex
Skill nameiced-rs
Stars
80
Forks
31
Bundled files
302
LicenseMIT
Links
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • 302 bundled files

    Scripts, templates, and references the model can read while it works. Files are read-only and never executed.

  • Open source

    Published by vanillagreencom on GitHub. Read the source before you install it.

Installation

Install the Iced Rs AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/vanillagreencom/kendex.git /tmp/kendex
mkdir -p .claude/skills
cp -r /tmp/kendex/skills/iced-rs .claude/skills/iced-rs
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Iced Rs in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads Iced Rs on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so Iced Rs is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

Iced 0.14

Workflow

  1. Classify the surface against references/guide-surface-selection.md. Do not skip.
  2. Read the canonical example in examples/, never generate 0.14 code from memory.
  3. Read the guide for that surface; for animated layered UI, references/guide-animated-layout.md comes first.
  4. Stuck: the guide's "Common failure modes" / "Gotchas", then references/guide-animation-debugging.md for animation and render bugs. The three most common: missing capture_event, missing invalidate_layout, 0.13 event signatures.

Surface choice: built-in widgets + .style(closure) for standard UI; Canvas for 2D custom drawing; Shader for GPU-dense rendering; iced::advanced::Widget for custom events, state, or layout; for a floating layer try tooltip, float, or stack+opaque before a custom Overlay.

Bundled resources

references/: API refs and guides

Full list in references/INDEX.md; load on demand.

GuideUse
guide-surface-selection.mdPick the right primitive
guide-custom-widgets.mdiced::advanced::Widget
guide-custom-overlays.mdiced::advanced::overlay::Overlay
guide-animated-layout.mdAnimated transitions, measured positions, keyed identity, clipping
guide-animation-debugging.mdSymptom→cause checklist for animation/render bugs
widgets.mdWidget catalog: every 0.14 widget, notes, canonical example

examples/: every upstream Iced 0.14 example

NeedRead first
Custom Widget implexamples/custom_widget/src/main.rs
GPU shader pipelineexamples/custom_shader/ (full dir)
Mesh / vector geometry widgetexamples/geometry/src/main.rs
Canvas 2D drawingexamples/bezier_tool, examples/clock, examples/color_palette
Canvas animationexamples/solar_system, examples/the_matrix, examples/game_of_life
Arc/ring animationexamples/loading_spinners, examples/arc
Modal dialogexamples/modal/src/main.rs (stack + opaque)
Toast/notification overlayexamples/toast/src/main.rs
Tooltip / zoom-on-hoverexamples/loupe/src/main.rs
Styled componentsexamples/styling/
pane_grid layoutexamples/pane_grid/
Multi-windowexamples/multi_window/
WebSocket subscriptionexamples/websocket/
Text editingexamples/editor/

iced_wgpu/ and external fallbacks

Renderer source for shader work: iced_wgpu/src/ (engine.rs, layer.rs, quad.rs, quad/solid.rs, quad/gradient.rs, triangle.rs, triangle/msaa.rs, primitive.rs, buffer.rs, shader/quad.wgsl). Local references are pinned to 0.14.0; prefer them. For newer API surface: ctx7 docs /websites/rs_iced_iced "<query>", https://docs.rs/iced/0.14.0/iced/, or upstream master at https://github.com/iced-rs/iced (may have unreleased APIs).

Breaking changes from Iced 0.13

  • Widget::update takes event: &Event (by ref, not by value)
  • Widget::layout takes &mut Tree
  • Keyboard subscriptions unified into keyboard::listen

Rules (non-negotiable framework invariants)

Widget tree consistency

Conditional wrapping changes tree shape and breaks event tracking. Always wrap; conditionally attach the handler.

rust
// WRONG: conditional wrapping changes tree shape
if dragging { mouse_area(label).into() } else { label.into() }

// RIGHT
let mut area = mouse_area(label);
if enable {
    area = area.on_press(msg);
}
area

MouseArea has no on_press_maybe (button-only); gate the on_press call, not the wrapper.

view() is pure

No side effects, no memoization dependent on call frequency. All mutable state lives in State and is mutated only in update(). Never trigger redraws from view().

Redraw vs rebuild

request_redraw() repaints but does not call view(). Animation state must live in widget::Tree state. Widget struct fields are frozen between view() calls. See references/animation.md § "Redraw vs rebuild."

Animation invalidation

Paint-only changes (color, opacity, rotation within fixed bounds) need shell.request_redraw(). Layout-affecting changes (size, position, expand/collapse, clipping bounds) need shell.request_redraw() and shell.invalidate_layout(). A widget that "only updates on the second click" has stale layout. Add invalidate_layout().

Draw order is z-order

In custom widget draw(), child iteration order determines z-order; last drawn is on top. stack semantics do not apply inside manual draw loops.

Overlay visibility requires layout invalidation

A widget that conditionally returns an overlay must call shell.invalidate_layout() when visibility changes.

Custom overlays are the #1 panic source

Prefer built-ins (tooltip, float, stack+opaque). A violated contract panics as container.rs unwrap() on None. The contract: children() returns a fixed count; diff() reconciles all children regardless of visibility; layout() returns nodes matching children; draw() walks the same tree layout produced. Full spec: references/guide-custom-overlays.md.

Overlay viewport contract

When calling descendant Widget methods from inside an Overlay impl, pass Rectangle::INFINITE as the viewport, never the stored viewport from the parent's overlay(). Overlay::layout() may still use bounds: Size for its own coordinate space.

Overlay state isolation

Overlay layers (stack children beyond the base) must not affect base-layer widget structure. Never change base-layer construction based on overlay presence.

Overlay starvation

Stacked mouse_area(...).interaction(...) layers can block underlying hover/move handlers even without opaque(...). Set Interaction::Grabbing on the real drag target, and reserve opaque(...) for true capture zones.

Hover stability

Hover sensors must not wrap content whose size changes during the animation they trigger. Use a stable outer hitbox. See references/guide-custom-widgets.md § "Stable hover hit regions."

Scroll state initialization

scrollable.on_scroll fires only after user scrolling, never at initial layout. Use sensor.on_show for initial layout and sensor.on_resize for changes.

Single message per interaction

One widget interaction produces one message. Composite actions (tab press becoming a drag) use a state machine in update(). When mouse_area handles semantics while button provides visual feedback, exactly one layer publishes: mouse_area owns the semantics and button stays visual-only.

button.on_press fires on mouse-up; mouse_area.on_press fires on mouse-down. Use it for drag initiation.

pane_grid

  • PaneGrid::min_size is uniform. Per-pane minimums must be enforced in pane content or in split/resize state.
  • TitleBar content must use Shrink width so empty space remains for the pick area; Fill eliminates it.
  • button and mouse_area both capture_event() on press: tab elements capturing means a custom tab drag, an empty title bar means native pane_grid drag.
  • Tab drag is mouse_area.on_press per tab plus listen_with for CursorMoved/ButtonReleased, with an Idle → Pressed(origin) → Dragging state machine at an 8px threshold.
  • In pane_grid::Content::update the title bar processes before the body. Do not unconditionally clear state in body-exit handlers that the title bar just established.
  • Keep drag feedback inside the picked pane subtree or in pane_grid::Style. mouse_area/opaque pane-drag overlays are rebuild-sensitive and can prevent Dropped events; drag previews must reuse the same TitleBar/body shell.

Subscriptions

Each data source needs stable identity: Subscription::run_with(id, stream) or .with(id), batched with Subscription::batch. See references/subscription.md.

Pre-aggregate high-frequency data in the subscription worker: emit one batch per non-empty ~16ms window, over bounded channels with producer-side try_send().

Theming: no custom Theme type for tokens

Build the palette with Theme::custom_with_fn("My Dark", palette, |p| theme::palette::Extended::generate(p)), keep app tokens in a LazyLock<AppTokens> sidecar, and route every visual value through it from style closures that ignore the passed &Theme. Introduce a custom Theme type only when runtime theme switching demands it.

Built-in palette roles are primary, success, danger, warning. Fonts load on the entry point (.font(include_bytes!(...))); Font::MONOSPACE resolves to the first loaded monospace font and Font::with_name("...") to a system font. See references/theme.md, references/theme-palette.md, references/catalog.md.

Cache staleness

Before writing cached or mirrored UI state, enumerate every mutation path that can stale it. Extend the existing global event path rather than adding a parallel subscription for the same event family; add at least one regression test per non-obvious invalidation or source-window gate.

Architecture

Message enum and State struct live in the root module; extracted modules receive &State or &mut State. Extract when a feature is gated and self-contained, forms a cohesive responsibility group, or exceeds ~30 lines over a well-defined State subset.

Multi-window: window::open(settings) -> Task<window::Id>, window::close(id). See references/window.md. Testing: iced_test provides Simulator (headless widget), Emulator (full runtime), and snapshot support.

Dev tools

cargo-hot, comet, the built-in F12 debugger, stress-test switches, and iced::debug::time: references/debug.md.

Bundled files

The model reads these on demand while the skill is loaded. They are exposed as readable files and are never executed.

and 140 more files.

Frequently asked questions

What does the Iced Rs AI skill do?

Load whenever building or debugging an Iced UI.

Why use Iced Rs on TypingMind?

Because you install it once and use it with any model. Iced Rs is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install Iced Rs in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/vanillagreencom/kendex/tree/main/skills/iced-rs. TypingMind reads its SKILL.md and bundles its files and installs it as a skill you can enable per chat.

Which AI models can use Iced Rs?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with Iced Rs?

As many as you like. As long as a model supports skills, you can use Iced Rs with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the Iced Rs AI skill free?

Yes. It is published on GitHub by vanillagreencom under the MIT license. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇