Welcome to FluxLoop

Ship Agents with Data. Scale Business.

🎯 Simulate at Scale

Run thousands of realistic multi-turn scenarios in parallel. Find edge cases before production.

📊 Align to Your Standards

Capture your implicit decision criteria. Turn intuition into automated evaluation.

🚀 Act on Insights

Reports that show what to fix and how. Analysis that drives action.

---

FluxLoop is an open-source toolkit for running reproducible, offline-first simulations of AI agents against dynamic scenarios. It empowers developers to rigorously test agent behavior, evaluate performance against custom criteria, and build confidence before shipping to production.

Why FluxLoop?

Building trustworthy AI agents requires systematic testing and evaluation. FluxLoop provides:

• Local-first simulation: Run experiments on your machine with full control
• Framework-agnostic: Works with any agent framework (LangGraph, LangChain, custom)
• Argument replay: Record complex function calls once, replay them hundreds of times
• Structured artifacts: Auditable JSON/JSONL outputs following a documented contract

Quick Start

Get started in minutes:

# Install packages
pip install fluxloop-cli fluxloop

# Initialize project
fluxloop init project --name my-agent
cd fluxloop/my-agent

# Generate inputs
fluxloop generate inputs --limit 50

# Run experiment
fluxloop run experiment

Documentation Structure

• Getting Started - Installation and setup
• Guides - Step-by-step tutorials
• SDK Reference - Python SDK documentation
• CLI Reference - Command-line tool documentation
• VSCode Extension - IDE integration guide

Core Concepts

🎯 Decorator-Based Instrumentation

Add @fluxloop.agent() to your agent function:

import fluxloop

@fluxloop.agent()
def run(input_text: str) -> str:
    return f"Response to: {input_text}"

🔄 Input Generation

Generate variations using LLM or deterministic strategies:

fluxloop generate inputs --limit 100

🧪 Offline Simulation

Run experiments locally without cloud dependencies:

fluxloop run experiment

📊 Structured Output

Every simulation produces auditable artifacts:

• summary.json - Aggregate statistics
• trace_summary.jsonl - Per-trace summary records
• observations.jsonl - Observation stream
• traces.jsonl - Detailed execution traces

What's Next?

Choose your path:

• New to FluxLoop? → Installation
• Ready to code? → Quick Start Guide
• Want examples? → Guides
• Need API reference? → SDK, CLI, VSCode

---

Ready to build trustworthy AI agents? Let's get started! 🚀