Base URL Configuration

Configure custom endpoints and base URLs for specific provider implementations, proxy setups, or alternative API endpoints.

Basic Base URL Configuration

Use the provider_base_url parameter to specify custom endpoints:

const client = new TokenlayOpenAI({
  provider_api_key: process.env.OPENAI_API_KEY,
  provider_base_url: "https://custom-proxy.example.com/v1",
  tokenlayKey: process.env.TOKENLAY_KEY,
});

Provider-Specific Base URLs

For multi-provider setups, specify base URLs for each provider:

const client = new TokenlayOpenAI({
  openai_key: process.env.OPENAI_API_KEY,
  openai_base_url: "https://openai-proxy.example.com/v1",
  
  anthropic_key: process.env.ANTHROPIC_API_KEY,
  anthropic_base_url: "https://anthropic-proxy.example.com/v1",
  
  tokenlayKey: process.env.TOKENLAY_KEY,
});

Common Use Cases

Corporate Proxy

Route through corporate firewalls or security proxies:

const client = new TokenlayOpenAI({
  provider_api_key: process.env.OPENAI_API_KEY,
  provider_base_url: "https://corporate-proxy.company.com/openai",
  tokenlayKey: process.env.TOKENLAY_KEY,
});

Azure OpenAI Service

Configure for Azure OpenAI endpoints:

const client = new TokenlayOpenAI({
  azure_key: process.env.AZURE_OPENAI_KEY,
  azure_base_url: "https://your-resource.openai.azure.com",
  tokenlayKey: process.env.TOKENLAY_KEY,
});

OpenRouter Integration

Use OpenRouter as your provider:

const client = new TokenlayOpenAI({
  provider_api_key: process.env.OPENROUTER_API_KEY,
  provider_base_url: "https://openrouter.ai/api/v1",
  tokenlayKey: process.env.TOKENLAY_KEY,
});

Self-Hosted Models

Connect to self-hosted or locally running models:

const client = new TokenlayOpenAI({
  provider_api_key: "local-key", // Can be any value for local models
  provider_base_url: "http://localhost:8000/v1",
  tokenlayKey: process.env.TOKENLAY_KEY,
});

URL Format Requirements

Base URLs should follow the OpenAI API format:

Include the version path (e.g., /v1)
Use HTTPS in production
Ensure the endpoint accepts OpenAI-compatible requests

// ✅ Correct format
"https://api.openai.com/v1"
"https://your-proxy.com/v1"
"https://localhost:8000/v1"
 
// ❌ Incorrect format  
"https://api.openai.com"        // Missing /v1
"http://prod-server.com/v1"     // HTTP in production
"https://non-openai-api.com/v2" // Wrong API format

Troubleshooting

Connection Issues

Verify your base URL is accessible:

curl -H "Authorization: Bearer $API_KEY" \
     https://your-base-url.com/v1/models

Authentication Errors

Ensure your API key works with the custom endpoint:

// Test with a simple models request
const response = await client.models.list();
console.log(response.data);

Proxy Configuration

For corporate environments, you may need additional proxy settings:

// Note: HTTP proxy settings depend on your environment
process.env.HTTPS_PROXY = "http://corporate-proxy:8080";

Provider Key Inference Model Parameter Behavior