Supported Providers

Apex AI Proxy supports multiple AI service providers, allowing you to aggregate their APIs behind a single unified interface. This page provides an overview of all supported providers and their capabilities.

Provider Overview

Provider	Models	Free Tier	Rate Limits	Special Features
Azure OpenAI	GPT-4, GPT-3.5, Embeddings	❌	High	Enterprise security, Custom deployments
DeepSeek	DeepSeek-Chat, DeepSeek-Coder	✅	Medium	High performance, Competitive pricing
Aliyun DashScope	Qwen, DeepSeek, GLM	✅	Medium	Multi-modal, Regional optimization
DeepInfra	Open source models	✅	Medium	Wide model selection, Fast inference

Quick Setup Comparison

Configuration Complexity

• Easiest: DeepInfra, DeepSeek
• Moderate: Aliyun DashScope
• Advanced: Azure OpenAI (requires Azure setup)

Best For Beginners

1. DeepSeek - Simple API, good free tier
2. DeepInfra - Many model options, straightforward setup
3. Aliyun DashScope - Good documentation, reliable service

Best For Production

1. Azure OpenAI - Enterprise features, high reliability
2. DeepSeek - Good performance and cost balance
3. Multi-provider setup - Maximum reliability through failover

Provider Selection Guide

🆓 Maximize Free Usage

If you want to take full advantage of free tiers:

const modelProviderConfig = {
  'general-chat': {
    providers: [
      { provider: 'deepseek', model: 'deepseek-chat' },      // Free tier
      { provider: 'deepinfra', model: 'meta-llama/Llama-2-7b-chat-hf' }, // Free tier
      { provider: 'aliyuncs', model: 'qwen-plus' },          // Free tier
    ],
  },
};

🏢 Enterprise Setup

For production environments with reliability requirements:

const modelProviderConfig = {
  'production-gpt4': {
    providers: [
      { provider: 'azure', model: 'gpt-4' },                 // Primary
      { provider: 'azure-backup', model: 'gpt-4' },          // Backup Azure
      { provider: 'deepseek', model: 'deepseek-chat' },      // Fallback
    ],
  },
};

🚀 High Performance

For applications requiring fast response times:

const modelProviderConfig = {
  'fast-chat': {
    providers: [
      { provider: 'deepseek', model: 'deepseek-chat' },      // Fast inference
      { provider: 'deepinfra', model: 'microsoft/DialoGPT-medium' }, // Fast model
    ],
  },
};

Detailed Provider Information

Azure OpenAI

• Best for: Enterprise applications, compliance requirements
• Strengths: High reliability, security features, dedicated resources
• Setup complexity: High (requires Azure account and resource setup)
• Cost: Pay-per-use, no free tier
• Detailed setup guide →

DeepSeek

• Best for: General purpose, cost-effective solutions
• Strengths: Good performance, reasonable pricing, simple API
• Setup complexity: Low (just need API key)
• Cost: Free tier available, competitive paid rates
• Detailed setup guide →

Aliyun DashScope

• Best for: Users in Asia-Pacific, multi-modal applications
• Strengths: Good regional performance, variety of models
• Setup complexity: Medium (requires Aliyun account)
• Cost: Free tier available, regional pricing
• Detailed setup guide →

DeepInfra

• Best for: Experimenting with open-source models
• Strengths: Wide model selection, good free tier
• Setup complexity: Low (simple API key setup)
• Cost: Generous free tier, pay-per-use
• Detailed setup guide →

Multi-Provider Configuration Example

Here's a comprehensive example showing how to configure multiple providers:

// wrangler-config.js
const providerConfig = {
  // Azure OpenAI
  azure: {
    base_url: 'https://your-resource.openai.azure.com/openai/deployments/your-deployment',
    api_keys: ['your-azure-api-key'],
  },
  
  // DeepSeek
  deepseek: {
    base_url: 'https://api.deepseek.com/v1',
    api_keys: [
      'your-deepseek-key-1',
      'your-deepseek-key-2',  // Multiple keys for higher limits
    ],
  },
  
  // Aliyun DashScope
  aliyuncs: {
    base_url: 'https://dashscope.aliyuncs.com/compatible-mode/v1',
    api_keys: ['your-aliyun-api-key'],
  },
  
  // DeepInfra
  deepinfra: {
    base_url: 'https://api.deepinfra.com/v1/openai',
    api_keys: ['your-deepinfra-api-key'],
  },
};

const modelProviderConfig = {
  // High-reliability GPT-4 equivalent
  'gpt-4': {
    providers: [
      { provider: 'azure', model: 'gpt-4' },
      { provider: 'deepseek', model: 'deepseek-chat' },
    ],
  },
  
  // Cost-optimized chat model
  'chat-model': {
    providers: [
      { provider: 'deepseek', model: 'deepseek-chat' },
      { provider: 'deepinfra', model: 'meta-llama/Llama-2-13b-chat-hf' },
      { provider: 'aliyuncs', model: 'qwen-plus' },
    ],
  },
  
  // Embeddings
  'text-embedding-ada-002': {
    providers: [
      { provider: 'azure', model: 'text-embedding-ada-002' },
      { provider: 'deepinfra', model: 'sentence-transformers/all-MiniLM-L6-v2' },
    ],
  },
};

module.exports = {
  providerConfig,
  modelProviderConfig,
  SERVICE_API_KEY: 'your-service-api-key',
};

Load Balancing Strategies

Round Robin (Default)

Requests are distributed evenly across all configured providers:

'balanced-model': {
  providers: [
    { provider: 'deepseek', model: 'deepseek-chat' },
    { provider: 'deepinfra', model: 'meta-llama/Llama-2-7b-chat-hf' },
    { provider: 'aliyuncs', model: 'qwen-plus' },
  ],
}

Priority-based Failover

Try providers in order, only use fallbacks when primary fails:

'reliable-model': {
  providers: [
    { provider: 'azure', model: 'gpt-4' },           // Primary
    { provider: 'deepseek', model: 'deepseek-chat' }, // Fallback 1
    { provider: 'deepinfra', model: 'meta-llama/Llama-2-13b-chat-hf' }, // Fallback 2
  ],
}

Monitoring and Optimization

Provider Performance Tracking

Monitor your providers to optimize configuration:

1. Response Time: Track which providers are fastest for your use case
2. Error Rates: Identify providers with reliability issues
3. Cost Analysis: Compare actual costs across providers
4. Rate Limit Usage: Monitor how close you are to limits

Optimization Tips

1. Use Multiple API Keys: Add multiple keys for the same provider to increase rate limits
2. Regional Selection: Choose providers with good performance in your region
3. Model Matching: Use similar capability models across providers for consistent failover
4. Cost Monitoring: Set up alerts for unexpected usage spikes

Getting Started

Quick Start (Recommended)

1. Start with DeepSeek for its simplicity and free tier
2. Add DeepInfra for additional free quota
3. Configure Azure OpenAI when you need enterprise features

For Production

1. Set up Azure OpenAI as your primary provider
2. Add DeepSeek as a high-quality fallback
3. Include Aliyun DashScope for additional redundancy

Next Steps

• Azure OpenAI Setup - Enterprise-grade OpenAI service
• DeepSeek Setup - High-performance models with free tier
• Aliyun DashScope Setup - Alibaba Cloud AI platform
• DeepInfra Setup - Open-source model inference
• Custom Provider Setup - Add your own providers

Need help choosing? Check our Configuration Guide for more detailed recommendations based on your specific use case.