How do I give ChatGPT, Claude, and Gemini the same memory?

Keep one record of yourself in plain files you own, and have every tool read it before it helps and save back what you tell it. I wrote my own system down as a blueprint, alexandria., and your own model reads it and builds yours on your computer, so you stop explaining yourself to every new chat.

Each app’s own memory knows only what happened in that app. If you only ever use ChatGPT, its memory may be all you need, and the gap opens the day you use a second tool or switch.

On your computer, Claude Code, Codex, and Cursor read and write the record directly, and each conversation in them is saved into it. ChatGPT, Claude, and Gemini read it through a copy they can open, such as one in your Google Drive, and a short block in each app’s saved instructions tells it to read your record first. If you only use a chat app, it can build the record in your Google Drive instead and save to it as you talk.

Each app keeps its own memory on, beside the record, and nothing you already have is moved or replaced. The record is plain files, so when a better model arrives it reads all of it on day one, and leaving costs you nothing.

Setup is one sitting. Your model takes a few minutes to look at how you work, shows you a short plan, and changes nothing until you say yes, and most of the build runs while you do something else.

Show the blueprint to the model you use most. It’ll tell you what you’re missing, if anything.