全球边缘LLM代理
毫秒级响应

// worker.js - LLM代理服务
export default {
  async fetch(request, env) {
    const url = new URL(request.url);
    
    // 验证API密钥
    const authHeader = request.headers.get('Authorization');
    if (!authHeader?.startsWith('Bearer ')) {
      return new Response('Unauthorized', { status: 401 });
    }

    // 解析请求体
    const body = await request.json();
    
    // 路由到目标模型
    const targetModel = body.model || 'gpt-3.5-turbo';
    const endpoint = getModelEndpoint(targetModel);
    
    // 转发请求
    const response = await fetch(endpoint, {
      method: 'POST',
      headers: {
        'Content-Type': 'application/json',
        'Authorization': `Bearer ${env.OPENAI_KEY}`
      },
      body: JSON.stringify(body)
    });
    
    return response;
  }
};

function getModelEndpoint(model) {
  const endpoints = {
    'gpt-4': 'https://api.openai.com/v1/chat/completions',
    'claude-3': 'https://api.anthropic.com/v1/messages',
    'gemini-pro': 'https://generativelanguage.googleapis.com/v1/models'
  };
  return endpoints[model] || endpoints['gpt-3.5-turbo'];
}

# wrangler.toml - Cloudflare Workers配置
name = "llm-proxy"
main = "worker.js"
compatibility_date = "2024-01-01"

# 环境变量
[vars]
ENVIRONMENT = "production"

# 密钥配置
[[kv_namespaces]]
binding = "CACHE"
id = "your-kv-namespace-id"

# 路由配置
[[routes]]
pattern = "api.yourdomain.com/*"
zone_name = "yourdomain.com"

# 限流配置
[limits]
cpu_ms = 50

# 环境配置
[env.production]
name = "llm-proxy-prod"
routes = [
  { pattern = "api.yourdomain.com/*", zone_name = "yourdomain.com" }
]

// 客户端调用示例
import OpenAI from 'openai';

const client = new OpenAI({
  baseURL: 'https://api.yourdomain.com/v1',
  apiKey: 'your-proxy-api-key'
});

// 发送请求
const response = await client.chat.completions.create({
  model: 'gpt-4',
  messages: [
    { role: 'user', content: 'Hello, how are you?' }
  ]
});

console.log(response.choices[0].message.content);

// 流式响应
const stream = await client.chat.completions.create({
  model: 'gpt-4',
  messages: [{ role: 'user', content: 'Tell me a story' }],
  stream: true
});

for await (const chunk of stream) {
  process.stdout.write(chunk.choices[0]?.delta?.content || '');
}