AI-Proxy-Worker/docs/Configuration.en.md

Configuration Guide

🌍 Language / 语言

🇺🇸 English | 🇨🇳 中文

AI Proxy Worker provides various configuration options to customize the proxy behavior according to your needs. Configuration is divided into two categories: environment variables and code configuration.

🔑 Environment Variables

These configurations are set through Cloudflare Workers' Secret feature, ensuring they remain invisible in the code for security.

Required Environment Variables

Variable	Type	Description	Example
DEEPSEEK_API_KEY	Secret	DeepSeek API key	sk-...

Setup Method:

wrangler secret put DEEPSEEK_API_KEY
# Enter your DeepSeek API key

Optional Environment Variables

Variable	Type	Description	Default
PROXY_KEY	Secret	Proxy access key for protecting your proxy	None (no authentication)

Setup Method:

wrangler secret put PROXY_KEY
# Enter a custom access key, e.g., my-secret-key-2025

Recommendations:

• Strongly recommended to set PROXY_KEY in production
• Use a strong password generator for access keys
• Rotate keys periodically for enhanced security

⚙️ Code Configuration

These configurations are defined in the CONFIG object in worker.js and can be modified as needed.

Core Configuration Options

DEEPSEEK_API_URL

• Type: String
• Default: 'https://api.deepseek.com/chat/completions'
• Description: DeepSeek API endpoint URL
• When to modify: Usually no need to change unless DeepSeek changes their API endpoint

DEEPSEEK_API_URL: 'https://api.deepseek.com/chat/completions'

MAX_BODY_SIZE

• Type: Number
• Default: 1024 * 1024 (1MB)
• Description: Maximum request body size limit (bytes)
• When to modify:
  • Increase: For longer conversation history
  • Decrease: For stricter limits to prevent abuse

MAX_BODY_SIZE: 1024 * 1024  // 1MB
MAX_BODY_SIZE: 2 * 1024 * 1024  // 2MB (more lenient)
MAX_BODY_SIZE: 512 * 1024  // 512KB (stricter)

REQUEST_TIMEOUT

• Type: Number
• Default: 30000 (30 seconds)
• Description: Timeout for requests to DeepSeek API (milliseconds)
• When to modify:
  • Increase: If frequently encountering timeout errors
  • Decrease: For faster failure responses

REQUEST_TIMEOUT: 30000   // 30 seconds (default)
REQUEST_TIMEOUT: 60000   // 60 seconds (more lenient)
REQUEST_TIMEOUT: 15000   // 15 seconds (stricter)

VALIDATE_REQUEST_BODY

• Type: Boolean
• Default: false
• Description: Whether to enable strict request body format validation
• When to modify:
  • true: Enable strict validation, checks messages format
  • false: Lenient mode, let DeepSeek API handle validation

VALIDATE_REQUEST_BODY: false  // Lenient mode (recommended)
VALIDATE_REQUEST_BODY: true   // Strict mode

DEFAULT_MODEL

• Type: String
• Default: 'deepseek-chat'
• Description: Default model to use when not specified in request
• When to modify: Based on your primary use case

DEFAULT_MODEL: 'deepseek-chat'      // General conversation (default)
DEFAULT_MODEL: 'deepseek-reasoner'  // Reasoning tasks primary

SUPPORTED_MODELS

• Type: Array
• Default: ['deepseek-chat', 'deepseek-reasoner']
• Description: List of supported models for validation and documentation
• When to modify: When DeepSeek releases new models

SUPPORTED_MODELS: [
  'deepseek-chat',      // General conversation model
  'deepseek-reasoner'   // Reasoning model
]

CORS Configuration

CORS_HEADERS

• Description: Cross-origin request response headers configuration
• Default: Allows all domains

const CORS_HEADERS = {
  'Access-Control-Allow-Origin': '*',
  'Access-Control-Allow-Methods': 'POST, GET, OPTIONS',
  'Access-Control-Allow-Headers': 'Content-Type, Authorization',
  'Access-Control-Max-Age': '86400',
};

Production Recommendations:

// Restrict to specific domains
'Access-Control-Allow-Origin': 'https://yourdomain.com'

// Or support multiple domains
'Access-Control-Allow-Origin': request.headers.get('origin') // Needs additional validation logic

Security Configuration

SECURITY_HEADERS

• Description: Security-related response headers
• Default: Basic security protection

const SECURITY_HEADERS = {
  'X-Content-Type-Options': 'nosniff',
  'X-Frame-Options': 'DENY',
  'X-XSS-Protection': '1; mode=block',
  'Referrer-Policy': 'strict-origin-when-cross-origin',
};

🛠️ Advanced Configuration

Custom Configuration Examples

For more complex configurations, you can modify like this:

// Development environment configuration
const CONFIG = {
  DEEPSEEK_API_URL: 'https://api.deepseek.com/chat/completions',
  MAX_BODY_SIZE: 2 * 1024 * 1024, // 2MB, more lenient
  REQUEST_TIMEOUT: 60000, // 60 seconds, longer timeout
  VALIDATE_REQUEST_BODY: true, // Enable strict validation
  DEFAULT_MODEL: 'deepseek-chat',
  SUPPORTED_MODELS: [
    'deepseek-chat',
    'deepseek-reasoner'
  ]
};

// Production environment configuration
const CONFIG = {
  DEEPSEEK_API_URL: 'https://api.deepseek.com/chat/completions',
  MAX_BODY_SIZE: 512 * 1024, // 512KB, stricter
  REQUEST_TIMEOUT: 15000, // 15 seconds, faster failure
  VALIDATE_REQUEST_BODY: false, // Lenient validation, better compatibility
  DEFAULT_MODEL: 'deepseek-chat',
  SUPPORTED_MODELS: [
    'deepseek-chat',
    'deepseek-reasoner'
  ]
};

Environment-Specific Configuration

You can also set environment-specific configurations in wrangler.toml:

name = "ai-proxy-worker"
main = "worker.js"
compatibility_date = "2025-08-17"

# Production environment
[env.production]
name = "ai-proxy-worker-prod"
vars = { ENVIRONMENT = "production" }

# Development environment
[env.development]
name = "ai-proxy-worker-dev"
vars = { ENVIRONMENT = "development" }

Then use in code:

// Adjust configuration based on environment
const isProduction = env.ENVIRONMENT === 'production';

const CONFIG = {
  // ... other configurations
  MAX_BODY_SIZE: isProduction ? 512 * 1024 : 2 * 1024 * 1024,
  REQUEST_TIMEOUT: isProduction ? 15000 : 60000,
  VALIDATE_REQUEST_BODY: !isProduction, // Enable strict validation in development
};

📝 Configuration Best Practices

1. Security Configuration

// ✅ Recommended: Restrict CORS in production
'Access-Control-Allow-Origin': 'https://yourdomain.com'

// ❌ Avoid: Using wildcards in production
'Access-Control-Allow-Origin': '*'

2. Performance Configuration

// ✅ Recommended: Reasonable timeout
REQUEST_TIMEOUT: 30000  // 30 seconds

// ❌ Avoid: Excessive timeout
REQUEST_TIMEOUT: 300000  // 5 minutes (too long)

3. Compatibility Configuration

// ✅ Recommended: Lenient validation mode
VALIDATE_REQUEST_BODY: false

// ⚠️ Note: Strict mode may cause some clients to fail
VALIDATE_REQUEST_BODY: true

🔧 Configuration Modification Steps

1. Modify Code Configuration

1. Edit worker.js file
2. Modify corresponding values in CONFIG object
3. Save file

2. Redeploy

wrangler publish

3. Verify Configuration

# Test health check
curl https://your-worker.workers.dev/

# Test API call
curl -X POST https://your-worker.workers.dev/chat \
  -H "Authorization: Bearer YOUR_PROXY_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model":"deepseek-chat","messages":[{"role":"user","content":"test"}]}'

❓ Common Configuration Questions

Q: How to increase request body size limit?

A: Modify MAX_BODY_SIZE configuration:

MAX_BODY_SIZE: 2 * 1024 * 1024  // Increase to 2MB

Q: How to set stricter timeout?

A: Modify REQUEST_TIMEOUT configuration:

REQUEST_TIMEOUT: 15000  // Reduce to 15 seconds

Q: How to disable request validation?

A: Set VALIDATE_REQUEST_BODY to false:

VALIDATE_REQUEST_BODY: false

Q: How to restrict CORS domains?

A: Modify Access-Control-Allow-Origin in CORS_HEADERS:

'Access-Control-Allow-Origin': 'https://yourdomain.com'

---

Need Help? 👉 Troubleshooting Guide | 📋 Issues