I’ve been using Empirical as my memory layer across AI tools.

Sam — Fri, 08 May 2026 16:36:24 +0000

ChatGPT memory helps.
Local MD files help.

But neither travels cleanly across everything I use, and packing too much into MD files eats context and tokens.

With Empirical, I keep my AGENTS.md lean and let Codex pull context dynamically when it actually needs it.

I can open ChatGPT on my phone, connected to Empirical, and it pulls the same memory context and writing tone I use in Codex or any other connected AI tool.
That means:

less repeated setup
cleaner, cheaper prompts
more consistent output across sessions

This is just the tip of the iceberg.

I wrote up a Codex example here:

How I Used Codex + Empirical to Lock In My Writing Voice | Empirical Blog

April 30 note on using Empirical with Codex to define a repeatable writing voice through guided questions and live revision.

empirical.gauzza.com

I Needed Memory That Survives Context Windows. Memory That Moves Across Environments

Sam — Thu, 09 Apr 2026 13:05:00 +0000

I kept running into the same thing with AI tools:

great context disappears
I repeat myself constantly
Every tool remembers different stuff (or nothing)
Moving between tools my context doesn't follow me

So I built Empirical.

It started in a pretty common place: I was iterating on a Philly-style hoagie roll recipe.

I wanted the AI to remember what I liked, what failed, and what I wanted to try next without re-explaining it every time.

I originally thought Empirical would be its own chatbot. I started down that path, then realized I was solving the wrong problem. Reinventing the wheel.

I didn’t need another chat interface.
I needed a memory layer I could use everywhere.

So I changed lanes and focused on MCP tools.

Now I use Empirical memory across:

Coding CLI's
ChatGPT
Claude Web
Claw Agents

Same memory, different interfaces. Now if ChatGPT is no longer _cool _ or Claude leaks it's entire codebase, I can switch to the latest hot thing and all my context and memories move with me.

Real examples that made this click for me

I can take a pic of a bourbon, say “I like this,” and that preference is saved as persistent memory.

I can send health data and query/chat over it later to help spot patterns.

I can write a PRD while going on a walk with ChatGPT, then pull it back up in a CLI session at my desk.

What’s next

I’m now working on connecting Empirical to more sources so memory reflects more of my actual life/workflow.

Current focus:

better pattern recognition over time
stronger multimodal memory (text + image + structured data)
cleaner memory workflows for agents

Forem: Sam