AI Agent Project with Memory and Tools
This project is a simple AI agent system build using Strands SDK, Ollama, and Mem0. It runs fully on a local machine ,so no external APIs are needed. the agent use tools ,remember past information and give smart responses using a local model.This project shows how to build a useful AI system in a simple and effective way for real world use. I learned the new things from this project .
This project was created for the AWS event From Zero to Agent Builder . the main goal is to build a smart AI agent using Stands SDK, Ollama, Mem0, and MCP. the system is fully local and does not require the APIs.
Challenge 1 is The Ai agent is think and give the interactive answer for the question asked by the user.
Challenge 2 is Adding tools and perform tasks . I added the tools such as Age Calculator , Calculator, Weather Assistant into the agent . The agent can answer the question like what's the weather in Chennai ? , How old is Someone born in 2007 ? , what is 2 ** 90 ? . At this stage, the agent works only with these tools.
Challenge 3 is Adds the persistent memory. the agent can remember past interactions, store user information , and retrieve it when needed , Users can also clear the stored memory. At this stage the agent does not use tools.
Challenge 4 is Combines both tools and memory to create a full AI agent . The agent can perform calculations, check weather, remember information , retrieve it, and delete it when the user requested. this stage is a combination of Challenge 2 and Challenge 3.
Challenge 5 is Introduces an MCP powered agent . it can handles the task like telling the current time , and date, and performing simple operation like reversing a words and calculator.
This project shows how tools can be used inside an AI system. the agent can choose the correct tool based on the user's request and complete the task.
Overall, this project is a simple and clear example of building a modern AI agent . I have explored this concept and gained practical knowledge by implementing it step by step.
Enjoyed reading this content? Let the author know!
Your likes, comments, shares, and saves help creators reach more builders.
Loading recommendations
Loading article