flint_client 0.0.4

A powerful, type-safe Dart client with HTTP, WebSocket, caching, retry, and AI provider support.

flint_client

Official Dart client for the Flint ecosystem.

flint_client gives you one package for:

• HTTP requests with typed parsing
• retry and cache support
• file upload and download helpers
• WebSocket communication
• AI provider integrations for OpenAI, Gemini, and Hugging Face

Installation

From pub.dev:

dart pub add flint_client

For Flutter:

flutter pub add flint_client

Or add it manually:

dependencies:
  flint_client: ^0.0.3+337

Quick Start

import 'package:flint_client/flint_client.dart';

Future<void> main() async {
  final client = FlintClient(
    baseUrl: 'https://api.example.com',
    debug: true,
  );

  final response = await client.get<Map<String, dynamic>>('/health');

  if (response.isSuccess) {
    print(response.data);
  } else {
    print(response.error?.message);
  }

  client.dispose();
}

Core Features

• Typed request and response handling
• Configurable retry policies
• In-memory caching with TTL controls
• Request and response lifecycle hooks
• Multipart uploads and file downloads
• WebSocket helpers through client.wc(...)
• AI providers with shared request flow and response helpers

AI Support

The package exports built-in AI providers so you can talk to model APIs using the same Flint request stack.

Available Providers

• OpenAIProvider
• GeminiProvider
• HuggingFaceProvider

AI Quick Start

import 'package:flint_client/flint_client.dart';

Future<void> main() async {
  final gemini = GeminiProvider(apiKey: 'YOUR_GEMINI_KEY');

  gemini.addContextMemory(
    'You are helping users understand the Flint Dart ecosystem.',
  );

  final response = await gemini.request(
    model: 'gemini-2.5-flash',
    prompt: 'Summarize what Flint Client does in one paragraph.',
  );

  final parsed = GeminiResponse.fromJson(response.data);
  print(parsed.text);
}

How To Use AI In flint_client

1. Create a provider with your API key.
2. Optionally add context with addContextMemory(...).
3. Call request(...) with a model name and prompt.
4. Parse the raw response with the matching response helper.
5. Reuse the same provider if you want conversation history preserved.

OpenAI Example

import 'package:flint_client/flint_client.dart';

Future<void> main() async {
  final openAI = OpenAIProvider(apiKey: 'YOUR_OPENAI_KEY');

  final response = await openAI.request(
    model: 'gpt-4o-mini',
    prompt: 'Write a short welcome message for Flint users.',
  );

  final parsed = OpenAIResponse.fromJson(response.data);
  print(parsed.text);
}

Gemini Example

import 'package:flint_client/flint_client.dart';

Future<void> main() async {
  final gemini = GeminiProvider(apiKey: 'YOUR_GEMINI_KEY');

  final response = await gemini.request(
    model: 'gemini-2.5-flash',
    prompt: 'Explain caching in simple terms.',
    includeHistory: true,
    includeContext: true,
    maxTokens: 300,
  );

  final parsed = GeminiResponse.fromJson(response.data);
  print(parsed.text);
}

Hugging Face Example

import 'package:flint_client/flint_client.dart';

Future<void> main() async {
  final hf = HuggingFaceProvider(apiKey: 'YOUR_HF_KEY');

  final response = await hf.request(
    model: 'gpt2',
    prompt: 'Generate a short API product tagline.',
  );

  final parsed = HuggingFaceResponse.fromJson(response.data);
  print(parsed.generatedText);
}

AI Notes

• AIProvider keeps in-memory history on the provider instance.
• resetHistory() clears conversation history.
• clearContextMemory() removes stored context snippets.
• includeHistory and includeContext let you control what gets sent.
• The raw provider response is still available through response.data.

HTTP Usage

GET With Parsing

final response = await client.get<User>(
  '/users/1',
  parser: (json) => User.fromJson(json),
);

POST JSON

final response = await client.post<Map<String, dynamic>>(
  '/products',
  body: {
    'title': 'New Product',
    'price': 29.99,
  },
);

Retry Configuration

final response = await client.get<String>(
  '/retry-test',
  retryConfig: RetryConfig(
    maxAttempts: 3,
    delay: Duration(seconds: 1),
    maxDelay: Duration(seconds: 10),
    retryStatusCodes: {500, 502, 503},
  ),
);

Cache Configuration

final response = await client.get<List<dynamic>>(
  '/products',
  cacheConfig: CacheConfig(
    maxAge: Duration(minutes: 10),
    forceRefresh: false,
  ),
);

File Upload

final response = await client.uploadFile<Map<String, dynamic>>(
  '/upload',
  fileField: 'image',
  file: File('path/to/image.jpg'),
);

WebSocket Usage

final ws = client.wc('/chat');

ws.on('connected', (_) => print('connected'));
ws.on('message', (data) => print(data));
ws.emit('send_message', {'text': 'Hello'});

Links

• Package page: https://pub.dev/packages/flint_client
• AI docs in this README: https://github.com/flint-dart/flint-client#ai-support
• Full docs: https://flintdart.eulogia.net/docs/client
• Repository: https://github.com/flint-dart/flint-client
• Examples: https://github.com/flint-dart/flint-client/tree/main/example
• AI example file: https://github.com/flint-dart/flint-client/blob/main/example/lib/main.dart
• Issue tracker: https://github.com/flint-dart/flint-client/issues

Development

dart pub get
dart test

License

This project is licensed under the MIT License. See LICENSE.