@haiec/openai

OpenAI SDK wrapper with automatic usage tracking and cost calculation for HAIEC AI Inventory.

Features

• ✅ Automatic Usage Tracking - Logs every API call to HAIEC
• ✅ Cost Calculation - Calculates costs based on model pricing
• ✅ Low Overhead - <10ms additional latency
• ✅ Error Handling - Gracefully handles tracking failures
• ✅ Drop-in Replacement - Works exactly like the official OpenAI SDK

Installation

npm install @haiec/openai openai

Usage

Basic Example

import TrackedOpenAI from '@haiec/openai';

const client = new TrackedOpenAI({
  apiKey: process.env.OPENAI_API_KEY!,
  haiecApiKey: process.env.HAIEC_API_KEY!,
});

const response = await client.chat.completions.create({
  model: 'gpt-4-turbo',
  messages: [{ role: 'user', content: 'Hello!' }],
});

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

Configuration

const client = new TrackedOpenAI({
  apiKey: 'sk-...', // Your OpenAI API key
  haiecApiKey: 'haiec_...', // Your HAIEC API key
  haiecEndpoint: 'https://haiec.com/api/v1/inventory/usage/log', // Optional: Custom endpoint
  organizationId: 'org-...', // Optional: OpenAI organization ID
});

Self-Hosted HAIEC

const client = new TrackedOpenAI({
  apiKey: process.env.OPENAI_API_KEY!,
  haiecApiKey: process.env.HAIEC_API_KEY!,
  haiecEndpoint: 'https://your-haiec-instance.com/api/v1/inventory/usage/log',
});

Supported Models & Pricing

| Model | Input Cost (per 1M tokens) | Output Cost (per 1M tokens) | |-------|---------------------------|----------------------------| | gpt-4-turbo | $10.00 | $30.00 | | gpt-4 | $30.00 | $60.00 | | gpt-4-32k | $60.00 | $120.00 | | gpt-3.5-turbo | $0.50 | $1.50 | | gpt-3.5-turbo-16k | $3.00 | $4.00 |

Pricing is automatically updated based on OpenAI's latest rates.

What Gets Tracked

For each API call, the following data is logged to HAIEC:

• Provider: openai
• Model: e.g., gpt-4-turbo
• Endpoint: e.g., chat.completions
• Request Tokens: Number of input tokens
• Response Tokens: Number of output tokens
• Cost: Calculated cost in USD
• Latency: Response time in milliseconds
• Status Code: HTTP status code
• Error Message: If the request failed

Performance

• Overhead: <10ms per request
• Async Logging: Usage is logged asynchronously to avoid blocking
• Error Handling: Tracking failures don't affect your API calls

Error Handling

If usage tracking fails, the error is logged to console but your OpenAI API call continues normally:

try {
  const response = await client.chat.completions.create({...});
  // Your response is returned even if tracking fails
} catch (error) {
  // Only OpenAI API errors are thrown
}

API Compatibility

This wrapper maintains 100% compatibility with the official OpenAI SDK. All methods and parameters work exactly the same.

Getting Your HAIEC API Key

1. Sign up at haiec.com
2. Navigate to Settings > API Keys
3. Click "Create API Key"
4. Copy your key and add it to your environment variables

Support

• Documentation: docs.haiec.com
• Issues: GitHub Issues
• Email: [email protected]

License

MIT