본문으로 건너뛰기

Quick Start

Get up and running with FluxLoop in 5 minutes.

Step 1: Create a Project​

Initialize a new FluxLoop project:

fluxloop init scenario --name my-agent
cd my-agent

This creates:

my-agent/
├── fluxloop.yaml          # Main configuration file
├── .env                   # Environment variables
├── src/
│   └── agent.py           # Sample agent code
└── scenarios/             # Test scenarios

Step 2: Instrument Your Agent​

Add the @fluxloop.agent() decorator to your agent function in src/agent.py:

import fluxloop

@fluxloop.agent()
def run(input_text: str) -> str:
    """Your agent logic"""
    # Process the input
    result = process_input(input_text)
    return result

Step 3: Authenticate​

Log in to the Web Platform:

fluxloop auth login
# Enter your API key from app.fluxloop.ai

Step 4: Define Scenarios​

Edit fluxloop.yaml to define your test personas and base inputs:

# fluxloop.yaml
personas:
  - name: novice_user
    description: A user new to the system
    
base_inputs:
  - input: "How do I get started?"
    expected_intent: help

Step 5: Generate Inputs​

Create diverse test case variations:

fluxloop generate --limit 50

This creates a local test bundle with 50 variations based on your personas.

Step 6: Run Test​

Execute your agent with the generated inputs:

fluxloop test

Results are saved locally in the results/ directory.

Step 7: View Results in the Cloud​

Upload and analyze your results on the Web Platform:

fluxloop sync upload

Open the provided link to results.fluxloop.ai to see:

  • Visual conversation traces
  • Performance metrics (latency, tokens, cost)
  • Success rate analysis

What You Get​

After running a test and uploading, you'll have:

  • Local Results - Structured JSONL data in ./results/
  • Cloud Dashboard - Interactive visualization and team collaboration
  • Performance Insights - Automatic analysis of agent behavior

Next Steps​

Congratulations! 🎉 You've run your first FluxLoop test.