Quick Start Guide

Get RunLog AI integrated with your agents in under 5 minutes.

Prerequisites

Python 3.8+

Node.js 16+

An existing AI agent or LLM application

Step 1: Installation

Install the SDK

Choose your preferred language

Python

pip install runlog-ai

TypeScript/JavaScript

npm install @runlog/sdk

Step 2: Basic Setup

Initialize RunLog

Add RunLog to your agent code

from runlog import RL

# Initialize RunLog
rl = RL(
    service="my-agent",
    env="development",
    api_key="your-api-key"  # Get from dashboard
)

# Wrap your agent logic
with rl.run(task="user-query"):
    # Your existing agent code here
    response = your_agent_function()
    return response

Step 3: Add Observability

Track Tool Calls

Monitor external API calls and tool usage

with rl.run(task="search-and-respond"):
    # Track tool calls
    results = rl.tool("search", query="user question")
    
    # Track LLM calls
    response = rl.llm("gpt-4", prompt=f"Answer: {results}")
    
    return response

Step 4: Add Safety Policies

Enforce Runtime Policies

Prevent unsafe actions before they happen

with rl.run(task="financial-query"):
    # Enforce budget limits
    rl.enforce("max_cost_per_run", limit=0.50)
    
    # Require approval for sensitive actions
    if "transfer" in user_input:
        rl.enforce("require_human_approval")
    
    response = process_query(user_input)
    return response

Next Steps

Continue learning about RunLog AI capabilities

Learn about Policies API Reference

Get Your API Key