Question 1

What is Letta?

Accepted Answer

Letta, formerly known as MemGPT, is a server for building AI agents with persistent long-term memory. Unlike a normal chatbot that forgets everything when a conversation ends, Letta manages structured memory in a PostgreSQL database so agents remember facts, past conversations, and accumulated knowl…

Question 2

Which port does it use?

Accepted Answer

This deployment runs Letta on host port 8285, mapped from the container's 8283. You build and manage agents through the Agent Development Environment, which you point at your server URL on that port.

Question 3

Do I need an OpenAI key, or can it run locally?

Accepted Answer

Either. Letta is model-agnostic — it handles memory and agent logic but delegates the actual thinking to an LLM provider you configure.

Question 4

How is its memory different from RAG?

Accepted Answer

RAG retrieves relevant chunks from a fixed knowledge base of documents you've loaded. Letta's memory accumulates dynamically from the agent's interactions — it remembers what you told it, how conversations went, and knowledge built up over time, managing what stays in context versus long-term storag…

Question 5

Where is the agent memory stored, and is it private?

Accepted Answer

In the PostgreSQL database this deployment runs, on your own server — which is why agent state survives restarts. That memory can be deeply personal, so keep the port internal and back up the database.

Port	Purpose
8285	Letta server and Agent Development Environment (mapped from container port 8283)
5432	PostgreSQL — internal to the Docker network, not exposed to the host

🤖 Setup Letta — Stateful AI Agents with Long-Term Memory

📦 Resources & Setup Scripts

Quick Install:

Tutorial Steps

1 Download the Script

2 Make it Executable

3 Run the Installer

4 Connect and Configure a Provider

Ports Used

Overview

Why Use It

When You Need It

Who Should Use It

Real Use Cases

Main Features

How to Use After Installation

Security Best Practices

Ports and Firewall Notes

Backup and Maintenance

Common Mistakes

Troubleshooting

Alternatives

When Not to Use It

PrismaTechWork Professional Help

Frequently Asked Questions