How I Implemented Memory in my AI Agent REXA

Description and use

<h1> How I Built Memory for REXA Using PostgreSQL and pgvector </h1> <p>One thing I wanted REXA to eventually have was <strong>long-term memory</strong>.</p> <p>The idea is simple: when I explicitly tell REXA something like:</p> <blockquote> <p>“Remember that I prefer PostgreSQL for my backend projects.”</p> </blockquote> <p>REXA should be able to save that information so it can be used later.</p> <p>The important distinction is that <strong>REXA does not currently retrieve these memories yet</strong>. Today, the implemented part is the <strong>memory-saving pipeline</strong>. Retrieval is the

Problem

<h1> How I Built Memory for REXA Using PostgreSQL and pgvector </h1> <p>One thing I wanted REXA to eventually have was <strong>long-term memory</strong>.</p> <p>The idea is simple: when I explicitly tell REXA something like:</p> <blockquote> <p>“Remember that I prefer PostgreSQL for my backend projects.”</p> </blockquote> <p>REXA should be able to save that information so it can be used later.</p> <p>The important distinction is that <strong>REXA does not currently retrieve these memories yet</strong>. Today, the implemented part is the <strong>memory-saving pipeline</strong>. Retrieval is the

Outcome

<h1> How I Built Memory for REXA Using PostgreSQL and pgvector </h1> <p>One thing I wanted REXA to eventually have was <strong>long-term memory</strong>.</p> <p>The idea is simple: when I explicitly tell REXA something like:</p> <blockquote> <p>“Remember that I prefer PostgreSQL for my backend projects.”</p> </blockquote> <p>REXA should be able to save that information so it can be used later.</p> <p>The important distinction is that <strong>REXA does not currently retrieve these memories yet</strong>. Today, the implemented part is the <strong>memory-saving pipeline</strong>. Retrieval is the

Evidence

How I Implemented Memory in my AI Agent REXA

Source