Build Agentic AI Apps with the Autogen framework

on YouTube ↗45m

" summary

AutoGen is an open-source framework that builds multi-agent LLM apps using AutoGen Studio (no-code) or the Python SDK (code-first). Examples show a GPT-4/GPT-4 Turbo assistant writing Python to scrape Wikipedia, retrying when the first run fails, and executing inside Docker for safety; skills are Python functions; workflows include a group chat manager that plans, assigns subtasks, and composes results (travel planner for Madrid/Barcelona, vegetarian options) and a pipeline that produced a GDP-per-capita plot. Community stats: ~20,000 Discord members, ~2,000 daily actives.

AI-written summary · about

" chapters (7)

1. 0:00Introduction and agenda
2. 1:31Understanding agentic AI
3. 4:48Overview of AutoGen
4. 7:32AutoGen Studio demo
5. 23:26Code-first development
6. 36:50Future designs and features
7. 42:46Conclusion and resources

" related talks

• 34mBuilding Multi-Agent Generative-AI Applications with AutoGenColombo et al. · PyData London 2024
• 33mAutoGen: A Multi-Agent AI FrameworkVictor Dibia, Chi Wang · AI.dev: Open Source GenAI & ML Summit North America 2023
• 25mFrontiers of Agentic AIChi Wang · Open Agent Summit 2025
• 10mKeynote: Next-Gen AI Applications via Multi-Agent ConversationVictor Dibia · AI.dev: Open Source GenAI & ML Summit North America 2023
• 10mKeynote: Next-Gen AI Applications via Multi-Agent ConversationVictor Dibia · Cassandra Summit 2023
• 20mGenAI agents - Evolution of automationArunansu Pattanayak · Global Artificial Intelligence Conference 2024
• 21mRevolutionizing software creation: Harnessing the power of GenAIMike Mason · ParadigmShift 2023
• 36mGenerative AI + Python: Unlocking Efficiency, Personalization, and InsightTony Ojeda · PyData Global 2024