Skip to main content
Empirical logoEmpirical

About

Empirical

A memory infrastructure company, and a small blog written by the people building it.

What Empirical is

Empirical is a persistent, queryable memory graph for AI tools. It stores project context, decisions, and durable notes once, and serves that context to ChatGPT, Claude, Cursor, Codex, and other coding agents through MCP and a CLI, so a fact recorded in one tool doesn't have to be re-explained to the next one.

Who publishes this site

Empirical is built and operated by Matt Gauzza, working independently. Posts on the blog are written by Matt, in first person, about real engineering work on this product, often produced alongside AI coding agents as part of that work.

Editorial standards

  • Posts describe real, verified events: incidents, fixes, decisions, and their actual outcomes, not hypothetical scenarios presented as fact.
  • Mistakes made along the way, including our own, are described honestly rather than edited out.
  • Technical claims are based on what we directly observed running our own systems, not secondhand reports.

Corrections

If something we published is factually wrong, tell us at hello@gauzza.com and we'll correct it. Substantive corrections are noted directly on the affected post.

Ownership and funding

Empirical is self-funded and independently owned. It is not sponsored content, and posts about our own product decisions are labeled as exactly that, our own account of our own work.