Sometime in the last year, without quite deciding to, you picked a way of telling your AI who you are.
Maybe you retype it every morning. Maybe you let ChatGPT quietly remember. Maybe you built a tidy Project in Claude once and never thought about it again. You chose a method, even if it never felt like choosing — and the method, more than the model, is what decides whether your AI still knows you tomorrow, or next week, or in the next app you open.
There are four of these methods, give or take. They are not equal. So let's put them side by side.
The four ways, side by side
| Method | Portable across tools? | Do you own it? | Effort | Best for |
|---|---|---|---|---|
| Re-explain every time | No — gone when the tab closes | Yes | High, every session | A one-off question |
| Let the tool remember | No — locked to that tool | Barely | None | Casual use in one app |
| Set up each tool | No — locked to that tool | Yes | Medium, repeated per tool | Power use in one tool |
| Portable context anchors | Yes — any tool | Yes | Low — write once, then maintain | Working across many tools |
The two columns that matter most are the first two: can you carry it everywhere, and does it actually belong to you. Hold onto those. They're what separate a method that scales from one that quietly traps you.
1. Re-explaining yourself every time
The default, and the one almost everyone starts with. You open a chat, type a few sentences about who you are and what you need, and get to work.
It's honest and it's flexible, and it has one fatal flaw: it evaporates. Close the tab and the context is gone, so tomorrow you do it again, and the day after that. You're not building anything — you're paying a small tax on every conversation, forever. Fine for a one-off. A slow leak if it's your whole system. For the longer version of why this wears you down, here's the case.
2. Letting the tool remember
Then the tools started remembering on their own. ChatGPT's memory notes things across your chats and pulls them forward when it decides they're relevant. No effort, mostly invisible, genuinely handy.
But it's the model's sketch of you, not your account of yourself — assembled from whatever you happened to mention, editable only through a little settings panel. And it stays put. What ChatGPT remembers never reaches Claude or Gemini or next year's tool. It's a memory that lives in one room of a house with a lot of rooms. I dug into exactly how this falls short if you want the full breakdown.
3. Setting up each tool
The power-user move: stop letting the model guess and hand it context on purpose. Claude Projects, custom GPTs, custom instructions — you pin who you are inside a workspace, and everything you do there starts from it.
Now you're in control, which is a real step up. But notice the shape of it. The context is bolted to that one tool. Build a beautiful setup inside a Claude Project and Claude is sharp; open ChatGPT and you're a stranger again. You did the work once, and the work stayed behind. Do this across four tools and you're not maintaining your context — you're maintaining four copies of it. And there's a subtler limit underneath all of it: a setup like this hands the tool instructions, not direction — the standing context of who you are.
4. Portable context anchors
The last method flips what the first three keep getting wrong. Instead of context that's stuck in a tool or stuck in a session, you write the durable facts about yourself once — who you are, how you sound, what you're building, who's in your orbit — into a plain document you keep, and you carry it to whatever model you're using that day.
That's a context anchor: structured text you own, that drops into any tool because it's just text. The tool changes. The context doesn't. You stop rebuilding yourself inside every new app, and the same you shows up in all of them. There are six anchors in all, each one a different durable piece of you, but the idea behind every one is the same — write it down, keep it, reuse it everywhere.
So which should you use?
Honestly, some of each. Let the tool's memory handle the small conveniences. Build a Project where you genuinely live inside one tool. Re-explain when it's a quick, throwaway ask.
But none of those is your foundation, because none of them travels and none of them is fully yours. The thing is, the method you want underneath all the others is the portable one — the context that's there no matter which tab you open, because you wrote it and you kept it. Get that right and the rest are just conveniences layered on top. Get it wrong, and you're back to being a stranger every time the tool changes.
Start with one
You don't have to build all of this today. You build the part that lasts, and you build it once.
So write a single anchor and see the difference for yourself. Start with the Personal Constitution — your values and beliefs, the foundation the rest stands on. It's free, it takes about thirty minutes, and the moment it's written it works in every tool you open. Not for one app. For all of them, and for whatever you're using a year from now.
Frequently Asked Questions
- What's the best way to give an AI context?
- Write portable context anchors. Of the four common methods — re-explaining yourself each session, platform memory, per-tool setups, and context anchors — anchors are the only one that is both portable across every tool and fully owned by you. You write your durable context once and reuse it in Claude, ChatGPT, Gemini, or any agent, instead of rebuilding it inside each one.
- Is ChatGPT memory enough to give my AI context?
- Not on its own. ChatGPT memory is convenient, but it's the model's running summary of you, it stays locked inside ChatGPT, and you only loosely control what it keeps. It won't follow you to Claude, Gemini, or a custom agent. For durable, portable context, write it down yourself in a context anchor you own.
- Do I have to set up context separately in every AI tool?
- Only if you rely on per-tool setups like Claude Projects, custom GPTs, or per-tool custom instructions — each one lives inside its own tool, so you rebuild your context every time you adopt a new app. A context anchor avoids that: it's a plain document you write once and paste or wire into any tool.
- What is a context anchor?
- A context anchor is a short, structured document that captures one durable part of who you are — your identity, voice, stories, situation, history, or relationships. Because it's plain structured text, it works in any AI tool, and because you write it, you own and control it.
- Which method should I start with?
- Start by writing a single context anchor — the Personal Constitution, which is free. It captures your values and beliefs, takes about thirty minutes, and works in every tool immediately. You can layer the other methods on top, but start with the one piece of context that travels with you everywhere.
