On owning your AI memory.
Plain-spoken guides and explainers on AI memory — what it is, why it should be yours, and how to keep the AI you've shaped across every model you use.
-
Why your local model forgets you — and how to give it a memory you own
Local models like Llama under Ollama or LM Studio start every session blank. Here's why your local AI forgets you, why that's actually fixable on your terms, and how to give it a persistent memory and persona you own.
Read -
MCP and the AI memory layer: how your AI reads one persona across every tool
MCP — the Model Context Protocol — is how a single persona reaches every editor and CLI you use. Here's what MCP is, how an AI memory layer uses it, and why it means one memory instead of four config files.
Read -
How to export your ChatGPT memory (and why a one-time copy isn't enough)
A step-by-step guide to exporting your ChatGPT memory — and why a one-time copy into the next app still leaves you without an AI memory you own. Here's the difference, and what to do instead.
Read -
Give your local AI model a memory that belongs to you
Local models like Llama under Ollama or LM Studio start every session blank. Here's how to give your local AI a persistent, local-first memory and persona you own — without anything leaving your machine.
Read -
ChatGPT memory vs. a portable AI memory you own
ChatGPT memory is convenient but locked to one app. Here's how it compares to a portable AI memory layer you own — across ownership, portability, the persona dimension, and privacy.
Read -
What is an AI memory layer, and why it matters in 2026
An AI memory layer holds your context and your AI's character in one portable place every assistant can read. Here's what that means, why vendor memory isn't enough, and what changed in 2026.
Read