MemGPT Agent FAQs

Question 1

How do I enable archival memory?

Accepted Answer

Activate the `pgvector` Spring profile. Ensure you have a compatible vector store and embeddings model configured. OpenAI and Postgres require minimal additional setup.

Question 2

What is MemGPT Agent?

Accepted Answer

MemGPT Agent is a Spring Boot-based implementation of a simplified MemGPT agent, exposed as a REST API or MCP server. It allows you to build and interact with LLMs using a memory-augmented architecture.

Question 3

What are the key features of MemGPT Agent?

Accepted Answer

Key features include core MemGPT functionality, REST API and MCP Server exposure, a command-line ChatBot, archival memory support via vector stores, and configurable agent instances through AgentManager and AgentLoader.

Question 4

What are the prerequisites for running MemGPT Agent?

Accepted Answer

You need Java 17 or later, Gradle, and an OpenAI API key. For archival memory, a vector store (like pgvector) and embeddings model are also required.

Question 5

How do I run the ChatBot application?

Accepted Answer

First, build either the REST or MCP server application. Configure the ChatBot application to target your chosen server in `application.yaml`. Finally, build and run the ChatBot application using Gradle.

MemGPT Agent

Características Principales

Casos de Uso

MemGPT Agent

Características Principales

Casos de Uso