Building an AI Flutter App in Minutes with Dart Genkit

AI is becoming a core part of app experiences, but integrating it into mobile apps often feels cumbersome. Setting up API clients, handling authentication, and managing network calls all add up. That is why Dart Genkit caught my attention.

In just a few minutes, I was able to create a fully functional AI app in Flutter that takes user prompts and generates responses using Google’s Gemini model. No complex setup, no boilerplate, just clean Dart code and a small Flutter UI.

Getting Started

Initializing Genkit is simple:

import 'package:genkit/genkit.dart';
import 'package:genkit_google_genai/genkit_google_genai.dart';

late final Genkit ai;

void initGenkit() {
  ai = Genkit(
    plugins: [
      googleAI(apiKey: dotenv.env['GEMINI_API_KEY'] ?? ''),
    ],
  );
}

Once initialized, sending a prompt is as easy as:

final result = await ai.generate(
  model: googleAI.gemini('gemini-2.5-flash'),
  prompt: 'What is Flutter?',
);
print(result.text);

That is it. No manual API requests, no parsing JSON responses, no boilerplate code.

Building the UI

I wrapped the AI functionality in a simple Flutter interface:

• A TextField to input prompts
• A FilledButton to generate responses
• Scrollable, selectable text area to display results
• Error handling and loading states

Here is a glimpse of how the UI handles the response:

if (_response.isNotEmpty)
  Expanded(
    child: Container(
      padding: const EdgeInsets.all(12),
      decoration: BoxDecoration(
        color: Colors.grey.shade100,
        borderRadius: BorderRadius.circular(8),
        border: Border.all(color: Colors.grey.shade300),
      ),
      child: SingleChildScrollView(
        child: SelectableText(
          _response,
          style: const TextStyle(fontSize: 15, height: 1.5),
        ),
      ),
    ),
  ),

The result is a responsive, clean interface that feels like a production-ready app even though it took minutes to build.

Why Dart Genkit Makes a Difference

1. Speed: You can go from empty project to working AI app in minutes.
2. Flexibility: Genkit supports multiple AI providers via plugins. You can easily swap models or services.
3. Clean API: The generate() function abstracts away the complexities of making requests and parsing responses.
4. Flutter-friendly: The library feels native to Dart and Flutter, keeping code readable and maintainable.

Taking It Further

This setup opens the door for many possibilities:

• AI-powered chatbots in apps
• Content generation tools
• Dynamic tutorials or onboarding inside apps
• Testing AI responses quickly during development

Because Genkit abstracts away most of the heavy lifting, you can focus on experimenting with AI interactions instead of worrying about setup.

Try It Yourself

I have made the full Genkit Playground app available on GitHub. It is fully functional and ready to run. Just add your Gemini API key and you are good to go.

In just a few minutes, Dart Genkit allowed me to explore AI capabilities in Flutter in a way that previously would have taken hours. If you are building AI-driven apps, it is worth giving it a try.