ai-cost-tracker 0.1.0

Zero-config OpenAI API cost tracking - just one import and you're done!

OpenAI Cost Tracker

Zero-config OpenAI API cost tracking - just one import and you're done!

Track all your OpenAI API usage automatically without modifying your existing code. Supports GPT-4, GPT-3.5, DALL-E, Sora, and more.

��� Features

• ��� Zero Configuration - Just one import line
• ��� Automatic Tracking - No code changes needed
• ��� Multi-Key Support - Track multiple API keys separately
• ��� Real-time Costs - See costs as they happen
• ��� SQLite Database - All data stored locally and securely
• ��� CLI Tools - Easy-to-use command-line interface
• ��� Daily Reports - Optional email summaries

���� Quick Start (2 Steps)

Step 1: Install

pip install ai-cost-tracker

Step 2: Add ONE Line

In your main Python file (the one that runs first):

import openai_cost_tracker.services.zero_config_tracker

That's it! All your OpenAI calls are now automatically tracked.

# Your existing code works as-is - no changes needed!
from openai import OpenAI

client = OpenAI()
response = client.chat.completions.create(
    model="gpt-4",
    messages=[{"role": "user", "content": "Hello"}]
)
# This call is automatically tracked! ���

���� View Your Costs

# View cost summary
openai-cost-view

# View last 7 days
openai-cost-view --days 7

# View by service
openai-cost-view --service my_service

# Export to JSON
openai-cost-view --export costs.json

���� Examples

Flask App

# app.py
from flask import Flask
import openai_cost_tracker.services.zero_config_tracker  # Add this line

app = Flask(__name__)
from openai import OpenAI
client = OpenAI()

@app.route('/chat')
def chat():
    # Automatically tracked!
    response = client.chat.completions.create(...)
    return response

FastAPI App

# main.py
from fastapi import FastAPI
import openai_cost_tracker.services.zero_config_tracker  # Add this line

app = FastAPI()
from openai import OpenAI
client = OpenAI()

@app.post("/chat")
async def chat():
    # Automatically tracked!
    response = await client.chat.completions.create(...)
    return response

Python Script

# my_script.py
import openai_cost_tracker.services.zero_config_tracker  # Add this line

from openai import OpenAI
client = OpenAI()

# All calls automatically tracked!
response = client.chat.completions.create(...)

���� Advanced Usage

Manual Tracking (Optional)

If you prefer manual tracking:

from openai_cost_tracker import get_tracker
from openai import OpenAI

client = OpenAI()
tracker = get_tracker()

response = client.chat.completions.create(...)
usage = response.usage

tracker.log_api_call(
    api_key=os.getenv("OPENAI_API_KEY"),
    api_type="chat",
    model="gpt-4",
    operation="create",
    cost=tracker.calculate_chat_cost(
        model="gpt-4",
        tokens_input=usage.prompt_tokens,
        tokens_output=usage.completion_tokens
    ),
    tokens_input=usage.prompt_tokens,
    tokens_output=usage.completion_tokens
)

Scan Your Codebase

Find all OpenAI usage in your project:

openai-cost-scan

# Scan specific directory
openai-cost-scan --scan-root /path/to/project

Daily Email Reports (Optional)

Set up daily email summaries:

# Install email support
pip install "ai-cost-tracker[email]"

# Configure email
mkdir -p ~/.openai_cost_tracker
cat > ~/.openai_cost_tracker/.env << EOF
SMTP_USER=[e-mail]
SMTP_PASSWORD=your-app-password
COST_REPORT_EMAIL=[e-mail]
EOF

# Run monitor
openai-cost-monitor

���� What Gets Tracked

• Chat Completions (GPT-4, GPT-3.5, etc.)
• Image Generation (DALL-E 2, DALL-E 3)
• Video Generation (Sora)
• Embeddings
• API Keys (hashed for privacy)
• Service Names (auto-detected)
• Tokens (input/output)
• Costs (calculated automatically)

���� Data Storage

All data is stored locally in SQLite:

• Default location: ~/.openai_cost_tracker/openai_usage.db
• Custom location: Set via CostTracker(db_path="/custom/path")

���� CLI Commands

openai-cost-view

View cost summaries and usage statistics.

openai-cost-view [--days N] [--service NAME] [--api-key HASH] [--recent N] [--json] [--export FILE]

openai-cost-scan

Scan codebase for OpenAI usage.

openai-cost-scan [--scan-root PATH] [--output FILE]

openai-cost-monitor

Run monitoring service with email reports.

openai-cost-monitor [--config PATH] [--db-path PATH] [--no-scan]

���� Pricing Support

Automatically calculates costs based on current OpenAI pricing:

• GPT-4: $30/$60 per 1M tokens (input/output)
• GPT-4o: $2.50/$10 per 1M tokens
• GPT-3.5-turbo: $0.50/$1.50 per 1M tokens
• DALL-E 3: $0.04-0.08/image
• Sora 2: $0.015/second

������� Configuration

Environment Variables

Optional environment variables:

# Enable verbose output
export OPENAI_COST_TRACKER_VERBOSE=1

# Custom database location
export COST_TRACKER_DB=/path/to/openai_usage.db

���� API Reference

CostTracker

from openai_cost_tracker import CostTracker, get_tracker

tracker = get_tracker()  # Get global tracker
# or
tracker = CostTracker(db_path="/custom/path")  # Custom tracker

# Calculate costs
cost = tracker.calculate_chat_cost("gpt-4", tokens_input=1000, tokens_output=500)
image_cost = tracker.calculate_image_cost("dall-e-3", size="1024x1024")
video_cost = tracker.calculate_video_cost("sora-2", seconds=8)

# Get summary
summary = tracker.get_usage_summary(days=7)

���� Contributing

Contributions welcome! Please see our contributing guidelines.

���� License

MIT License - see LICENSE file

���� Support

• Issues: GitHub Issues
• Documentation: See docs in repository

---

Made with ������ for developers who want to track their AI costs without the hassle.