Platform Support

Understanding platform requirements and compatibility for expo-ai-kit.

Overview

expo-ai-kit provides on-device AI capabilities by leveraging native platform frameworks. Currently, iOS is fully supported through Apple's Foundation Models framework, while Android support is in development.

iOS Support

Requirements

iOS To use expo-ai-kit on iOS, you need:

• iOS 26.0 or later — The Foundation Models API was introduced in iOS 26.0
• Apple Intelligence enabled — Must be turned on in Settings > Apple Intelligence & Siri
• Compatible hardware — See supported devices below
• Xcode 16+ — Required for building

Supported Devices

Apple Intelligence requires specific hardware with Neural Engine capabilities:

Checking Availability

Always check availability at runtime, as users may be on unsupported devices or have Apple Intelligence disabled:

import { isAvailable } from 'expo-ai-kit';
import { Platform } from 'react-native';

async function checkSupport() {
  // Quick platform check
  if (Platform.OS !== 'ios') {
    return { supported: false, reason: 'iOS only' };
  }

  // Check Apple Intelligence availability
  const available = await isAvailable();

  if (!available) {
    return {
      supported: false,
      reason: 'Apple Intelligence not available. Check device compatibility and settings.',
    };
  }

  return { supported: true };
}

Android Support

Android Beta Android support is currently in development.

We're exploring integration with:

• Google ML Kit — For on-device ML features
• Gemini Nano — Google's on-device LLM for supported devices
• MediaPipe LLM Inference — Cross-device LLM support

In the meantime, you can prepare your codebase by properly checking availability:

import { isAvailable } from 'expo-ai-kit';

// This will return false on Android until support is added
const available = await isAvailable();

if (!available) {
  // Use cloud AI fallback or show appropriate message
}

For Android-specific setup instructions (when available), see the Android Setup Guide.

Web Support

Web support is not planned for expo-ai-kit.

The library's core value is on-device AI processing, which requires native platform APIs. Web browsers don't provide equivalent APIs for running LLMs locally.

For web applications, consider:

• Cloud-based AI services (OpenAI, Anthropic, Google AI)
• WebGPU-based solutions for experimental local inference
• Progressive enhancement: use cloud AI on web, on-device on mobile

Graceful Degradation

Design your app to handle cases where on-device AI isn't available. Here's a recommended pattern:

import { isAvailable, createSession, sendMessage } from 'expo-ai-kit';

type AIProvider = 'on-device' | 'cloud' | 'none';

export function useAI() {
  const [provider, setProvider] = useState<AIProvider>('none');

  useEffect(() => {
    async function detectProvider() {
      const onDeviceAvailable = await isAvailable();

      if (onDeviceAvailable) {
        setProvider('on-device');
      } else if (CLOUD_AI_ENABLED) {
        setProvider('cloud');
      } else {
        setProvider('none');
      }
    }
    detectProvider();
  }, []);

  const sendAIMessage = async (message: string) => {
    switch (provider) {
      case 'on-device':
        const session = await createSession();
        const response = await sendMessage(session, { message });
        await session.close();
        return response.text;

      case 'cloud':
        return await cloudAI.complete(message);

      default:
        throw new Error('No AI provider available');
    }
  };

  return { provider, sendMessage: sendAIMessage };
}