Study real implementations, then adapt them to your own product
Every example page states what the code demonstrates, which principles it supports, which libraries it uses, and how it fits into the learning and certification path.
Key Facts
- Examples
- 103
- Runnable
- 103 runnable scripts
- Browse by
- Principle, runtime pattern, difficulty, library, and product problem
- Libraries
- pydantic, python-dotenv, openai, requests, docling, gradio, streamlit, ipykernel
- Use this for
- Learning patterns, comparing implementations, and preparing your own product decisions
Agentic Design Patterns
The shapes agents are built from, read against the ten principles.
Every pattern states which principles it expresses in code, which its textbook form breaks, and what to add. That third column is the part nobody else publishes.
Explore all patternsDry-Run: simulate before you execute
Any agent that can take an action the user would not want undone by accident: deletion, payment, sending, deploying, writing to production. Use it wherever 'the agent did it before I could stop it' is a plausible incident report.
Episodic and semantic memory: remember turns, and remember facts
Any assistant expected to know the user across sessions. The split matters most once the transcript is long: loading old conversations to 'remember' someone is expensive and drags the answer toward whatever was being discussed then.
Works with
In this section
Providers & libraries
Anthropic
OpenAI
Google Gemini
Mistral AI
Meta Llama
CursorCourse stage
Module
Model context
From agent demos to runtime discipline
A capable model is not a runtime architecture. If agents are going to trigger workflows, load files, use tools, delegate work, and act across channels, the runtime needs clear patterns for control, visibility, and recovery. This cluster helps teams design those patterns deliberately.