ai_prompt_kit 0.0.2

A Flutter-friendly AI helper package for calling LLM APIs using prompt templates and structured responses.

ai_prompt_kit 🤖✨

A Flutter-friendly AI helper package for calling LLM / AI APIs using prompt templates, clean abstractions, and safe response handling.

This package is designed to make AI integration in Flutter simple, readable, and production-ready.

---

✨ Features

• 🔌 OpenAI-style REST API support
• 🧠 PromptTemplate system (variable-based prompts)
• 📦 Clean AI client abstraction
• ⚠️ Graceful error handling (no crashes)
• 🎯 UI-agnostic (works with any Flutter app)
• 🚀 pub.dev–ready example app

---

📦 Installation

Add this to your pubspec.yaml:

dependencies:
  ai_prompt_kit: ^0.0.1

Then run:

flutter pub get

---

🚀 Quick Start

1️⃣ Initialize the AI Client

final aiClient = AiClient(
  AiConfig(
    apiKey: "YOUR_API_KEY",
    baseUrl: "https://api.openai.com/v1",
  ),
);

⚠️ Do not hardcode API keys in packages.
Pass the key from your app or environment.

---

2️⃣ Simple Prompt Completion

final response = await aiClient.complete(
  prompt: "Explain Flutter packages in simple words",
);

if (response.hasError) {
  print(response.error);
} else {
  print(response.text);
}

---

🧠 PromptTemplate (Key Feature)

PromptTemplate allows you to create reusable prompts using variables instead of string concatenation.

Example: Summarization Prompt

final prompt = PromptTemplate(
  template: "Summarize the following text in {lang}:\n{text}",
  variables: {
    "lang": "English",
    "text": "Flutter packages help developers share reusable components.",
  },
);

final response = await aiClient.run(prompt);

if (response.hasError) {
  print(response.error);
} else {
  print(response.text);
}

✔️ Why use PromptTemplate?

• Cleaner prompt construction
• Easy variable replacement
• Reusable & maintainable prompts
• Ideal for production AI systems

---

📊 AI Response Model

All AI calls return an AiResponse object.

class AiResponse {
  final String text;
  final int? tokens;
  final String? error;

  bool get hasError => error != null;
}

---

⚠️ Error Handling (No Crashes)

Handled cases:

• Missing API key
• Invalid API key (401)
• Network failures
• Invalid or empty AI responses

Example error message:

API key not configured. Please provide a valid API key.

---

🧪 Example App

This package includes a full Flutter example demonstrating:

• PromptTemplate usage
• AI request lifecycle
• Loading state handling
• Error-safe UI

Run the example:

cd example
flutter run

---

🔐 API Key Management

For security reasons:

• ❌ Do not commit API keys to GitHub
• ❌ Do not hardcode API keys in packages
• ✅ Provide API keys from the application layer
• .env usage is optional (not required)

---

🛣️ Roadmap

• Streaming AI responses
• Structured JSON output
• Multiple LLM provider support
• Built-in prompt presets

---

🤝 Contributing

Contributions are welcome!
Feel free to open issues or submit pull requests.

---

📄 License

MIT License