Poorgramer-Zack/dart-expert-skills / skills/flutter-genui
generating-flutter-ui
VerifiedGenerative UI (GenUI) for Flutter with the genui 0.10 package (SurfaceController, A2uiTransportAdapter, Conversation, Surface) and an LLM backend such as Firebase AI Logic (firebase_ai). Use when the layout must come from model output, for chat-based UI generation, prompt-to-UI systems, or AI-driven dashboards. Do not use for static screens.
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 5.7 KB |
SKILL.md
View on GitHub---
name: "generating-flutter-ui"
description: "Generative UI (GenUI) for Flutter with the genui 0.10 package (SurfaceController, A2uiTransportAdapter, Conversation, Surface) and an LLM backend such as Firebase AI Logic (firebase_ai). Use when the layout must come from model output, for chat-based UI generation, prompt-to-UI systems, or AI-driven dashboards. Do not use for static screens."
metadata:
last_modified: "2026-11-10"
package_versions: "genui 0.10.4, firebase_ai 5.1.0, firebase_core 5.1.0 (verified with flutter analyze on Flutter 3.47.9)"
---
# Flutter Generative UI (GenUI)
## Goal
Render UI that an AI model generates at runtime, using `genui` 0.10. Use it only for features whose structure depends on model output. Static screens should use normal widgets.
## Instructions
### 1. Setup
```bash
flutter pub add genui firebase_ai firebase_core
```
- macOS and iOS: allow outbound network access in `{ios,macos}/Runner/*.entitlements` with `com.apple.security.network.client` set to `true`.
- Initialize Firebase and configure GenUI logging before `runApp`. `DefaultFirebaseOptions` comes from the FlutterFire CLI.
```dart
Future<void> main() async {
WidgetsFlutterBinding.ensureInitialized();
await Firebase.initializeApp(options: DefaultFirebaseOptions.currentPlatform);
configureLogging(logCallback: (level, msg) => debugPrint('GenUI $level: $msg'));
runApp(const MaterialApp(home: GenUiPage()));
}
```
### 2. Core objects
| Object | Role |
|---|---|
| `SurfaceController(catalogs: [...])` | Holds the widget catalog and the state of each surface. |
| `A2uiTransportAdapter(onSend: ...)` | Sends user messages to the model. Feed model output back with `addChunk(text)`. |
| `Conversation(controller:, transport:)` | Runs the loop. Exposes `events` and `sendRequest(ChatMessage)`. |
| `Surface(surfaceContext: controller.contextFor(id))` | Renders one surface. |
### 3. Example: Firebase AI Logic backend
This file compiles with `flutter analyze` against the pinned versions.
```dart
import 'package:firebase_ai/firebase_ai.dart';
import 'package:flutter/material.dart';
import 'package:genui/genui.dart' hide TextPart;
import 'package:genui/genui.dart' as genui;
class GenUiPage extends StatefulWidget {
const GenUiPage({super.key});
@override
State<GenUiPage> createState() => _GenUiPageState();
}
class _GenUiPageState extends State<GenUiPage> {
late final SurfaceController _controller;
late final A2uiTransportAdapter _transport;
late final Conversation _conversation;
late final ChatSession _chatSession;
final List<String> _surfaceIds = [];
@override
void initState() {
super.initState();
final catalog = BasicCatalogItems.asCatalog();
_controller = SurfaceController(catalogs: [catalog]);
_transport = A2uiTransportAdapter(onSend: _sendAndReceive);
_conversation = Conversation(controller: _controller, transport: _transport);
final model = FirebaseAI.googleAI().generativeModel(model: 'gemini-3-flash-preview');
_chatSession = model.startChat();
_conversation.events.listen((event) {
switch (event) {
case ConversationSurfaceAdded added:
setState(() => _surfaceIds.add(added.surfaceId));
case ConversationSurfaceRemoved removed:
setState(() => _surfaceIds.remove(removed.surfaceId));
case ConversationContentReceived content:
debugPrint(content.text);
case ConversationError error:
debugPrint('${error.error}');
default:
}
});
final promptBuilder = PromptBuilder.chat(
catalog: catalog,
systemPromptFragments: ['You are a helpful assistant.'],
);
_conversation.sendRequest(ChatMessage.system(promptBuilder.systemPromptJoined()));
}
Future<void> _sendAndReceive(ChatMessage msg) async {
final buffer = StringBuffer();
for (final part in msg.parts) {
if (part.isUiInteractionPart) {
buffer.write(part.asUiInteractionPart!.interaction);
} else if (part is genui.TextPart) {
buffer.write(part.text);
}
}
if (buffer.isEmpty) return;
final response = await _chatSession.sendMessage(Content.text(buffer.toString()));
if (response.text?.isNotEmpty ?? false) {
_transport.addChunk(response.text!);
}
}
@override
Widget build(BuildContext context) {
return Column(children: [
for (final id in _surfaceIds) Surface(surfaceContext: _controller.contextFor(id)),
]);
}
@override
void dispose() {
_conversation.dispose();
super.dispose();
}
}
```
### 4. Events
Handle `ConversationSurfaceAdded` and `ConversationSurfaceRemoved` by updating the list of surface IDs, then build one `Surface` per ID. Handle `ConversationContentReceived` for plain text and `ConversationError` for failures.
### 5. System prompt
Build the prompt with `PromptBuilder.chat(catalog:, systemPromptFragments:)`, then send `ChatMessage.system(promptBuilder.systemPromptJoined())` once. `systemPrompt()` is a method that returns an `Iterable<String>`, not a `String`.
### 6. Ask about input widgets
Before you write user-input UI, ask which widgets the app should use. The official guide treats this as a question for the user, not a default.
### 7. Official reference
The `genui` package ships an official integration guide at `skills/integrate-genui-firebase/SKILL.md` inside the installed package. Read it when you need the full Firebase setup steps.
## Constraints
- Do not add `genui_firebase_ai`. It is discontinued. Call `firebase_ai` directly from `onSend`.
- Do not use the 0.7-era names `ContentGenerator`, `GenUiConversation`, `A2uiMessageProcessor`, or the manual `DataModel` wiring. They do not match genui 0.10.4.
- Dispose the `Conversation` in `dispose()`.
- Use GenUI only for AI-driven structure. Do not use it for static screens or ordinary layouts.
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| brainstormingobra/superpowersYou MUST use this before any creative work - creating features, building components, adding functionality, modifying behavior, or planning anything new, in software or out of it (a talk, a business, a renovation). | Stars: 323k | Updated | Claude CodeCodex | Verified | Install |
| brand-guidelinesanthropics/skillsApplies Anthropic's official brand colors and typography to any sort of artifact that may benefit from having Anthropic's look-and-feel. Use it when brand colors or style guidelines, visual formatting, or company design standards apply. | Stars: 198k | Updated | Claude CodeCodex | Verified | Install |
| frontend-designanthropics/skillsGuidance for distinctive, intentional visual design when building new UI or reshaping an existing one. Helps with aesthetic direction, typography, and making choices that don't read as templated defaults. | Stars: 198k | Updated | Claude CodeCodex | Verified | Install |
| mcp-builderanthropics/skillsGuide for creating high-quality MCP (Model Context Protocol) servers that enable LLMs to interact with external services through well-designed tools. Use when building MCP servers to integrate external APIs or services, whether in Python (FastMCP) or Node/TypeScript (MCP SDK). | Stars: 198k | Updated | Claude CodeCodex | Verified | Install |