DeepSeek API

Validate Pub Version License: MIT

A modern Flutter package for seamless integration with DeepSeek's powerful AI APIs. Supports all major features, including chat completions, model management, and streaming.

Installation

Add the following dependency to your pubspec.yaml:

dependencies:
  deepseek_api: ^0.1.1

Then run:

flutter pub get

Initialization

import 'package:deepseek_api/deepseek_api.dart';

final deepseek = DeepSeekAPI(
  apiKey: 'your-api-key',
  // Optional: baseUrl: 'https://api.deepseek.com/v1' (default)
);

🔗 Get API Key | 💵 Pricing Details

Quick Start

Basic Chat Completion

final response = await deepseek.createChatCompletion(
  ChatCompletionRequest(
    model: 'deepseek-chat',
    messages: [
      ChatMessage(
        role: 'user',
        content: 'Explain quantum computing in 50 words'
      )
    ],
    temperature: 0.7,
    maxTokens: 100,
  ),
);

print(response.choices.first.message.content);

Streaming Responses

final response = await deepseek.createChatCompletion(
  ChatCompletionRequest(
    model: 'deepseek-chat',
    messages: [/* your messages */],
    stream: true,
  ),
);

// Handle stream using Dio's response stream

Supported Models

Latest Models (2024)

Model ID Description Best For
deepseek-chat General-purpose chat model Everyday conversations
deepseek-coder Code-focused model Programming assistance
deepseek-math Advanced mathematical reasoning Complex calculations
deepseek-chat-pro Premium chat model Enterprise applications

🔍 Full Model List | 🔄 Model Updates

Parameters Reference

ChatCompletionRequest

Parameter Type Default Description
model String Required The model ID to use
messages List Required Conversation history as ChatMessage objects
temperature double? 1.0 Creativity control (0.0-2.0)
maxTokens int? null Maximum number of tokens to generate
topP double? 1.0 Nucleus sampling threshold (0.0-1.0)
frequencyPenalty double? 0.0 Penalize new tokens based on frequency (-2.0 to 2.0)
presencePenalty double? 0.0 Penalize new tokens based on presence (-2.0 to 2.0)
stop List? null Up to 4 sequences where the API will stop generating
stream bool? false Whether to stream back partial progress
user String? null Unique identifier for end-user

Advanced Features

Model Management

// List all available models
final models = await deepseek.listModels();
print(models.data.map((m) => m.id).join(', '));

Error Handling

try {
  // API call
} on RateLimitException catch (e) {
  print('Slow down! ${e.message}');
} on ApiException catch (e) {
  print('API Error: ${e.message}');
} on NetworkException {
  print('Check your internet connection');
}

Custom Configuration

final customDio = Dio()
  ..interceptors.add(LogInterceptor());

final deepseek = DeepSeekAPI(
  apiKey: 'your-key',
  baseUrl: 'https://api.deepseek.com/v2', // Custom endpoint
);

API Resources

Contributing

We welcome contributions! Please:

  1. Fork the repository
  2. Create your feature branch
  3. Submit a PR with tests

📌 Report Issues | 📖 View Source

Disclaimer: This package is not officially affiliated with DeepSeek. Always refer to the official API documentation for the most accurate information.

Libraries

deepseek_api