Mobile AI Integration: Calling Chinese LLMs in iOS/Android Apps

Mobile Security Considerations

Never hardcode API Keys in apps — they'll be extracted via decompilation. Correct approach: use a backend proxy or server-side signed temporary tokens.

Flutter Integration

import 'package:http/http.dart' as http;
import 'dart:convert';

Future<String> chat(String prompt) async {
  final res = await http.post(
    Uri.parse('https://chinawhapi.com/v1/chat/completions'),
    headers: {
      'Authorization': 'Bearer {server_token}', // server-side token
      'Content-Type': 'application/json',
    },
    body: jsonEncode({
      'model': 'qwen3.5-flash',
      'messages': [{'role': 'user', 'content': prompt}],
    }),
  );
  return jsonDecode(res.body)['choices'][0]['message']['content'];
}

React Native Integration

Use the OpenAI SDK (React Native supports it) or call the API directly with fetch. Using a backend proxy is recommended for additional security.

Network Optimization

Mobile networks are unreliable — implement: request timeout (30-60 seconds recommended), auto-retry (up to 2 times), offline queue (store requests when offline, retry when connection restores).