One of the biggest advantages of building AI agents over traditional automations is that they are supposed to act more like a human. But for being so human, they sure have a pretty terrible memory most of the time.
We want our agents to get smarter over time through interacting with us - remembering goals, instructions, corrections, our preferences, etc.
This kind of long term memory is when you really start to take your agents to the next level of personalization and human like behavior. In this video, I’ll show you step by step how to build these self-learning AI agents using an open source Python library called Mem0 which is specifically built for this purpose.
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
All the Mem0 agent code covered in this video can be found here:
https://github.com/coleam00/ottomator-agents/tree/main/mem0-agent
Mem0:
https://github.com/mem0ai/mem0
Visit the oTTomator Live Agent Studio and try out the Mem0 agent for yourself for free!
https://studio.ottomator.ai/
You get 100 tokens when you sign up, and tokens are just my way to make sure you don’t drain my bank account testing my agents ;)
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
00:00 - Self-Evolving Agents with Mem0
01:35 - Using Gemini to Show We Need Mem0
03:02 - Pydantic AI Mem0 Agent Demo
04:48 - Mem0 GitHub (100% Open Source)
05:29 - Our Mem0 Implementation Overview
06:44 - Version 1 - Basic Mem0 Setup
10:24 - Version 1 Demo
12:09 - Version 2 - Mem0 + Supabase Integration
13:23 - Version 2 Demo
15:53 - Version 3 - Frontend + Supabase Authentication
17:44 - Version 3 Demo
21:26 - Bonus Section - How Mem0 Works
24:31 - Outro
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Join me as I push the limits of what is possible with AI. I'll be uploading videos at least two times a week - Sundays and Wednesdays at 7:00 PM CDT!