Don’t Miss This CrewAI Update to Transform YOUR AI AGents

---

---

AI Summary

Video Summary: Crew AI Updates and Code Examples

Crew AI Callbacks

• New before kickoff and after kickoff callbacks introduced.
• Update Crew AI to the newest version.
• Use crew AI create crew command to see new decorators in the crew base.
• before kickoff is an optional hook executed before the crew starts.
• after kickoff is called after the crew finishes and outputs results.
• Example provided showing how to print inputs and manipulate data before it’s sent to the crew.

Knowledge Support

• Support added for enriching agents with Knowledge (capital K).
• Knowledge allows agents to access structured data sources like PDFs, spreadsheets, etc.
• Example shown using sampl knowledge.py to demonstrate how to give knowledge to an agent.
• Knowledge is stored in a chroma database for efficient retrieval.

User Preferences and Memories with Mem Zero

• Mem Zero (mem0) support for retrieving user preferences and memories.
• Previous video covered mem0 with Autogen.
• Example code provided for integrating mem0 with Crew AI using chat.
• Demonstrated how to create a chatbot that uses mem0 to store and retrieve user preferences.
• Showed how to use mem0 to give personalized recommendations based on stored memories.

Additional Updates

• Async execution fixes.
• Upgrades to CLI Watson supported models.
• Reduced level for Bandit.
• Fixed some tests and updated documentation.

Resources and Community

• Mention of a school community and a Crew AI Master course with weekly updates.
• Introduction to using n8n and Crew AI together.
• Free gift offered in the no introduction course.

GitHub and Course Links

• Code examples are available on GitHub.
• Links to the Crew AI Master course and additional resources provided in the video description.

(Note: No specific CLI commands, website URLs, or detailed instructions were provided in the transcript text.)