Cookbook¶

You're in Step 5 of the Learning Path

You've learned the patterns. Now grab these complete, working examples that you can copy, modify, and use in production.

These recipes are production-ready implementations that demonstrate LionAGI patterns in real-world scenarios. Each includes full code, expected outputs, performance metrics, and customization guidance.

Available Recipes¶

Analysis & Research¶

Claim Extraction - Extract and validate claims from documents
Research Synthesis - Aggregate multiple sources into insights

Business Applications¶

HR Automation - Multi-agent HR workflow system
Code Review Crew - Parallel code analysis with quality gates

Creative Work¶

Brainstorming - Generate and refine ideas collaboratively

Technical¶

Data Persistence - Save agent state to databases

Quick Templates¶

Basic Multi-Agent Analysis¶

from lionagi import Branch, iModel
import asyncio

agents = {
    "analyst": Branch(system="Analyze data", chat_model=iModel(provider="openai", model="gpt-4o-mini")),
    "critic": Branch(system="Find issues", chat_model=iModel(provider="openai", model="gpt-4o-mini")),
    "advisor": Branch(system="Give recommendations", chat_model=iModel(provider="openai", model="gpt-4o-mini"))
}

async def analyze(topic):
    results = await asyncio.gather(*[
        agent.chat(f"Analyze: {topic}") 
        for agent in agents.values()
    ])
    return dict(zip(agents.keys(), results))

Sequential Pipeline¶

from lionagi import Session, Branch, Builder

async def pipeline(input_data):
    session = Session()
    builder = Builder("pipeline")

    extract = builder.add_operation("chat", instruction=f"Extract key points from: {input_data}")
    analyze = builder.add_operation("chat", depends_on=[extract], instruction="Analyze the extracted points")
    summarize = builder.add_operation("chat", depends_on=[analyze], instruction="Create executive summary")

    return await session.flow(builder.get_graph())

Contributing Examples¶

Have a useful pattern? Submit a PR with:

Working code
Clear use case
Expected output
Performance metrics

Congratulations! 🎉

You've completed the LionAGI learning path! You now have:

✅ Understanding of LionAGI's paradigm and advantages
✅ Knowledge of core concepts and architecture
✅ Proven patterns for common workflows
✅ Production-ready examples to build upon

What's next? - Advanced Topics - Custom operations, performance tuning, observability - Integrations - Connect with databases, tools, and services - Migration Guides - If you're coming from other frameworks