ai_prompt_kit 0.1.1

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-compatible REST API support\
• 🧠 PromptTemplate system (variable-based prompts)\
• 🌍 Multi-language support via AiLanguage
• 🔎 Auto language detection\
• 📊 Token usage tracking\
• ⚠️ Graceful error handling (no crashes)\
• 🎯 UI-agnostic\
• 🚀 Production-ready example app

---

📦 Installation

Add to your pubspec.yaml:

dependencies:
  ai_prompt_kit: ^0.1.1

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",
  ),
);

⚠️ Never hardcode API keys in production apps.
Use environment variables or a backend proxy.

---

🧠 Using PromptTemplate

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

final response = await aiClient.run(prompt);

if (response.hasError) {
  print(response.error);
} else {
  print(response.text);
  print("Tokens used: ${response.tokens}");
}

---

🌍 Language Control

Force response language:

language: AiLanguage.bengali
language: AiLanguage.chinese
language: AiLanguage.arabic

Enable automatic detection:

await aiClient.complete(
  prompt: userInput,
  autoDetectLanguage: true,
);

---

📊 AI Response Model

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

  bool get hasError => error != null;
}

---

⚠️ Error Handling

Handled scenarios:

• Missing API key\
• Invalid API key (401)\
• Network failures\
• Invalid AI response\
• Unexpected API errors

---

🔐 Security Best Practice

Recommended architecture:

Flutter App → Your Backend → OpenAI

Instead of:

Flutter App → OpenAI directly

---

🛣️ Roadmap

• 🌊 Streaming responses\
• 🧾 Structured JSON output mode\
• 🧠 System + User role messaging\
• 🔄 Multi-provider support\
• 📦 Built-in AI presets

---

👨‍💻 Project Maintainer ❤️

Md. Hamid Hosen
Associate Software Engineer @P2M Soft

---

📄 License

MIT License