Quick Start Guide

Get started with LLMGovernor in under 5 minutes. This guide will walk you through setting up cost tracking for your AI agents.

Prerequisites

• Python 3.8 or higher
• An active AI agent project using OpenAI, Anthropic, or other supported providers
• A LLMGovernor account (sign up at dashboard.Vivekreddy99.com)

Installation

Install the LLMGovernor Python SDK using pip:

pip install llmgovernor

Configuration

1. Get your API key from the LLMGovernor dashboard

2. Set your environment variable:

export LLMGOVERNOR_API_KEY="your-api-key-here"

Or add it to your .env file:

LLMGOVERNOR_API_KEY=your-api-key-here

Basic Usage

Add the @fleet.agent decorator to any function that makes LLM API calls:

import openai
from blaze_agents import fleet

# Set your OpenAI API key
openai.api_key = "your-openai-key"

@fleet.agent(budget=50.00)  # Set $50 budget
def my_research_agent(query: str):
    """Research agent with cost tracking."""
    response = openai.chat.completions.create(
        model="gpt-4",
        messages=[
            {"role": "system", "content": "You are a helpful research assistant."},
            {"role": "user", "content": query}
        ],
        max_tokens=500
    )
    return response.choices[0].message.content

# Use your agent
result = my_research_agent("What are the latest developments in AI safety?")
print(result)

Advanced Configuration

Per-User Budgets

Track costs per user or session:

@fleet.agent(budget=25.00, user_id="user_123")
def personalized_agent(user_input: str, user_id: str):
    # Agent logic here
    pass

Custom Tags

Add custom tags for better cost categorization:

@fleet.agent(
    budget=100.00,
    tags={"department": "marketing", "campaign": "q4_2024"}
)
def marketing_agent(content: str):
    # Agent logic here
    pass

Budget Enforcement

When a budget is exceeded, LLMGovernor will raise a BudgetExceededError:

from blaze_agents.exceptions import BudgetExceededError

try:
    result = my_research_agent("Complex query requiring many tokens")
except BudgetExceededError as e:
    print(f"Budget exceeded: {e}")
    # Handle gracefully - perhaps use a cheaper model
    result = fallback_agent("Simplified query")

Monitoring

Visit your LLMGovernor dashboard to:

• View real-time cost tracking
• Set up alerts and notifications
• Analyze spending patterns
• Configure budget rules

Multi-Provider Support

LLMGovernor automatically detects and tracks costs for:

• OpenAI (GPT-3.5, GPT-4, etc.)
• Anthropic (Claude models)
• Google (PaLM, Gemini)
• Azure OpenAI
• Cohere
• Hugging Face

No additional configuration needed - just use the decorator!

Next Steps

• Check out the SDK Reference for advanced features
• Explore the API Documentation for custom integrations
• Join our Discord community for support

Need Help?

• 📚 Browse our documentation
• 💬 Join our Discord community
• 📧 Email us at [email protected]
• 🐛 Report issues on GitHub

---

Ready to track your AI costs? Sign up for free and get 100 events per month at no cost.