An animated, touch-reactive bot avatar for native Android (Jetpack Compose) and iOS (SwiftUI). A fluffy mascot with 17 moods, drawn natively: no web view, no Lottie, no video files. Use it as the face of an AI assistant, chatbot, voice assistant, onboarding guide or loading state, and recolour it or change its mood at runtime.
- Native on both platforms. Jetpack Compose
Canvason Android, SwiftUICanvason iOS. Same design, same motion constants, same API shape on both. - 17 moods. Happy, Thinking, Idea, Love, Listening, Speaking, Success, Error and more. Switch with one parameter and the face crossfades.
- Made for assistant UIs. Listening, Thinking and Speaking states map directly onto a voice or chat assistant's turn-taking.
- Touch reactions built in. Tap to giggle, press and hold for love.
- Fully styleable. Any colour, fur length and density, brightness, six built-in body shapes, plus your own shape from an SVG path.
- Custom faces. Mix 7 eye styles and 10 mouth styles to build any expression.
- Dependency-light. Plain Kotlin and Swift. Android needs only AndroidX Compose, iOS only SwiftUI.
- Respects accessibility. Animation stops when the system reduces motion.
| Part | Options |
|---|---|
| Body shapes (6) | Teardrop (default), Round, Egg, Blob, Blossom, Squircle |
| Custom body | any SVG path via BotShape.svg(...) (the demo apps add Bean and Heart this way) |
| Eyes (7) | Oval, ArcUp, ArcDown, Heart, Star, Cross, Slit |
| Mouths (10) | None, Smile, Grin, Open, O, Flat, Frown, Wavy, Dot, Talk |
| Face mode (2) | Eyes, EyesAndMouth |
| Moods (17) | Neutral, Happy, Working, Sleeping, Thinking, Idea, Love, Sad, Surprised, Angry, Confused, Excited, Listening, Speaking, Success, Error, Giggle |
| Badges | thought bubble (Thinking), light bulb (Idea) |
| Style knobs | colour, brightness, fur length, fur density |
| Face knobs | eyeScale, eyeScaleRight, eyeTilt, lookX, lookY, mouthScale, mouthShift |
| Touch | tap = giggle, press and hold = love |
minSdk 24, Jetpack Compose. Published through JitPack:
// settings.gradle.kts
dependencyResolutionManagement { repositories { maven("https://jitpack.io") } }
// app/build.gradle.kts
implementation("com.github.knowamit:interactive-bot-ui:0.2.0")Or include the module directly: copy android/botui and add include(":botui").
iOS 16+, SwiftUI. In Xcode use File > Add Package Dependencies and enter
https://github.com/knowamit/interactive-bot-ui, or in Package.swift:
.package(url: "https://github.com/knowamit/interactive-bot-ui", from: "0.2.0")Then import BotUI.
import dev.botui.*
BotView(
state = BotState.Happy, // see States below
modifier = Modifier.size(260.dp),
style = BotStyle(
color = Color(0xFFF5292A),
brightness = 1.1f, furLength = 2.6f, furDensity = 0.75f,
face = BotFace.EyesAndMouth, // or Eyes
),
)Module android/botui. Demo app: android/app. Build: cd android && ./gradlew :app:assembleDebug.
Tests: :botui:testDebugUnitTest.
import BotUI
BotView(
state: .happy, // see States below
style: BotStyle(color: Color(red: 0.96, green: 0.16, blue: 0.16),
brightness: 1.1, furLength: 2.6, face: .eyesAndMouth))
.frame(width: 260, height: 260)Swift package at the repo root. Demo app: ios/Demo (xcodegen generate, then open the project).
Tests: swift test at the repo root.
Both demos show every body shape in one horizontal row (four visible at a time, swipe for the rest) and drive them all from the same controls: mood, colour, brightness, fur length, mouth and custom face. Tap a bot to giggle, hold for love.
Map your assistant's phases to states:
| Assistant phase | State |
|---|---|
| Idle | Neutral |
| Listening to the user | Listening |
| Waiting for the model | Thinking |
| Speaking a reply | Speaking |
| Got a good idea | Idea |
| Task finished | Success |
| Something failed | Error |
BotView(state = when (phase) {
Phase.Listening -> BotState.Listening
Phase.Thinking -> BotState.Thinking
Phase.Speaking -> BotState.Speaking
else -> BotState.Neutral
})17 presets, same names on both ports (Swift uses lowerCamel):
Neutral, Happy, Working, Sleeping, Thinking, Idea, Love, Sad, Surprised, Angry, Confused, Excited, Listening, Speaking, Success, Error, Giggle.
Each state is a face preset (BotFaceSpec) plus a body motion (hop, shake, nod, heartbeat, tilt,
giggle wiggle). Changing state crossfades the old face into the new one over about 0.22s. No face
uses eyebrows. Sad and Angry tilt the eyes instead (eyeTilt).
Thinking shows a thought bubble with pulsing dots, and Idea shows the system light bulb emoji with pulsing rays, both floating beside the head.
Tap the bot to make it giggle (1.6s), press and hold for love (2.4s). Then it returns to the state you set.
BotView(state, touchReactions = true, onTouch = { reacted -> /* Giggle or Love */ })BotView(state: state, touchReactions: true, onTouch: { reacted in /* .giggle or .love */ })Pass touchReactions = false to turn it off.
Built-in shapes: Teardrop (default), Round, Egg, Blob, Blossom, Squircle.
style = BotStyle(shape = BotShape.Round)
style = BotStyle(shape = BotShape.svg("M20 40C20 10 80 10 80 40C95 55 85 95 50 95C15 95 5 55 20 40Z"))BotStyle(shape: .round)
BotStyle(shape: .svg("M20 40C20 10 80 10 80 40C95 55 85 95 50 95C15 95 5 55 20 40Z"))svg takes the first sub-path of any SVG path string (M L H V C S Q Z, absolute or relative, any
units, no arcs). It is scaled to fit and stands on the ground. A bad path falls back to the teardrop.
Fur and shading follow the shape. The demo apps also include Bean and Heart shapes made this way.
Build any state by passing your own face through the style. A custom face replaces the state's face
(motion still follows state).
style = BotStyle(customFace = BotFaceSpec(
eyes = BotEyes.Star, mouth = BotMouth.Open, eyeScale = 1.2f, eyeTilt = -0.4f))BotStyle(customFace: BotFaceSpec(
eyes: .star, mouth: .open, eyeScale: 1.2, eyeTilt: -0.4))Eye and mouth names are listed in What is included (Swift spells the empty mouth
.none). eyeTilt runs -1 (angry) to 1 (sad) and turns oval and slit eyes.
The bot has no built-in margin on Android, so badges such as the light bulb and the hop float past the top of the body. Leave some room above the head (the demo uses about 24dp).
Both ports share the same sim constants (seeded mulberry32, hop, blink, gaze), body path and shading. Fur is a close approximation, not a pixel match across platforms.
Issues and pull requests are welcome. Run the tests before opening a PR: swift test at the root and
cd android && ./gradlew :botui:testDebugUnitTest.
MIT, see LICENSE.
