https://github.com/sushant1827/long-term-memory-agent

Build an AI Agent with long term memory using LangGraph
https://github.com/sushant1827/long-term-memory-agent

agentic-workflow email-assistant episodic-memory gpt-4o-mini langchain langgraph semantic-memory tools

Last synced: about 1 month ago
JSON representation

Build an AI Agent with long term memory using LangGraph

• Host: GitHub
• URL: https://github.com/sushant1827/long-term-memory-agent
• Owner: sushant1827
• Created: 2025-03-26T15:41:23.000Z (about 1 year ago)
• Default Branch: main
• Last Pushed: 2025-03-30T17:01:56.000Z (about 1 year ago)
• Last Synced: 2025-04-06T22:13:17.663Z (about 1 year ago)
• Topics: agentic-workflow, email-assistant, episodic-memory, gpt-4o-mini, langchain, langgraph, semantic-memory, tools
• Language: Jupyter Notebook
• Homepage:
• Size: 1.32 MB
• Stars: 0
• Watchers: 1
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

          
# Baseline Email Assistant

Let's create an Email Assistant that:

- Classifies incoming messages (respond, ignore, notify)

- Drafts responses

- Schedules meetings

![Screenshot 2025-03-26 160025](https://github.com/user-attachments/assets/18cb60b0-1426-4e9c-b992-db0aceda735d)

---

# Email Assistant with Semantic Memory

Let's create an Email Assistant that:

- Classifies incoming messages (respond, ignore, notify)

- Drafts responses

- Schedules meetings

Now, we'll add memory and give the assistant the ability to remember details from previous emails.

![image](https://github.com/user-attachments/assets/a6be7d4a-5367-43d9-b2f4-07f82f71e736)

---

# Email Assistant with Semantic Memory + Episodic Memory

Let's create an Email Assistant that:

- Classifies incoming messages (respond, ignore, notify)

- Drafts responses

- Schedules meetings

- Uses memory to remember details from previous emails

Now, we'll add human-in-the-loop following the triage step to better refine the assistant's ability to classify emails

![image](https://github.com/user-attachments/assets/2f3a753f-9d4b-4e80-bd97-3afbfd944a0a)

---

# Email Assistant with Semantic + Episodic + Procedural Memory

We previously built an email assistant that:

- Classifies incoming messages (respond, ignore, notify)

- Uses human-in-the-loop to refine the assistant's ability to classify emails

- Drafts responses

- Schedules meetings

- Uses memory to remember details from previous emails 

Now, we'll add procedural memory that allows the user to update instructions for using the calendar and email writing tools. 

![image](https://github.com/user-attachments/assets/1e23aa0e-021b-4be4-aef0-38da2784ffcf)

---