So, I was having a cup of coffee with my friend Alex last week and he says to me, Elliot, I'm still using chat GPT, but everyone keeps talking about how much better Claude is. And to be fair, Claude are killing it right now. They've got Opus 4.6. They've got Cowwork, they've got Claude Code, and if you're a business, then it's hands down the number one choice right now. So, yeah, I asked him like, "Why don't you just switch to Claude then?" He says, "Well, I've spent 6 months in chat GPT. has got all my work and my chat memory and I really don't want to have to start everything again. He's built too much in chat GPT to walk away from. And if you're in a similar position, then I've got you because you can migrate everything from chat GPT to Claude in less than 5 minutes. And I'm going to show you exactly how to do it. So, let me show you my exact migration system that will take you just a few minutes. It's inside my AI recipe vault here. And here is the recipe here that we're going to be following. You can get a link to everything here, including all the prompts and step-by-step instructions in the description below. But the first thing we need to do, obviously, is export information out of Chat GPT. And this is really easy to do. In the bottom left of your chat GPT screen here, you click your little settings button and then you want to go into data controls. Inside data controls, right at the bottom, you've got export data. Now, all you have to do is hit export and it's going to ask you a couple of questions here. It's going to give you all your account details, chats to be included in the export. It's going to be sent to your registered email. So, make sure you can access the email that you signed up to chat GPT with. It's going to give you a download link that will expire within 24 hours. And it might take a little time depending on how much data you've got in there to actually produce. So, you might have to wait a few minutes for the email to actually arrive. But, we're just going to hit confirm export. So, you should get an email like this that says your data export is ready. And then all you got to do is hit download data export. Once you've downloaded it, then we can start to move into a platform like Claude. So this is the normal Claude home screen. And what I want to do from here, I actually want to get this information inside a specific project. So I'm going to go to my project and I'm going to create a new project. I've actually created one earlier called chat GPT hub, but I'll walk you through this process again. So we'll call this chat GPTHub 2 and hit create project. Now from here, we're going to be able to add information. And this is one way of doing it. If you have started using Claude and you're using Claude co-work then I'm going to go into an even easier way to do it with Claude co-work in a minute but this is the easiest way to get started. The beauty of projects in chat GPT and Claude is we can add files to use as our knowledge base in this project. So we can hit the plus button here and click upload from device. Then when you downloaded your data you should have got a file that looks something like this. A zip file with a long string of letters and numbers. You need to obviously open up that zip file and what it's going to give you is all the information from your chat GPT including your chat history and any files you've uploaded and downloaded from there. So what we want to do this chat html we want to actually open this up in claude or upload it into our project knowledge base in claude. So once we upload that you might see a problem. If you've only got a small chat history, then it should be fine. But we can only add up to 31 megabytes of chat history into our clawed projects. So that presents a problem. But I do have some workarounds for this if that is happening to you. The first one is to go and open up that HTML file. So we can go back to that data export file on our computer and doubleclick and open up chat.html. That's going to literally open up a chat window with all of that chat information that we've ever had with chat GPT. And what you can do from here is you can obviously read through it and see all your chats in one window, but you can hit command A and actually copy paste all of this text. So we can copy all of that text in this window and actually go back into our clawed project, hit add files, but this time you want to click add text content because we can just go and paste this text content straight into this window. So, we'll paste that text in there. And also, I'll just call this chat GPT chat history there. And we can add that content. It might take a little while. Now, we're ready to rock and roll. We can actually access our chat GPT memory and history within this project if we just want to keep it siloed in this project. I'm just going to show you how to transplant that into your full clawed workspace in a second. But essentially, if we say, "What do you know about Elliot Prince?" It's going to go and read that knowledge file to give us our answer from all of our chat GPT history. You can see we've got you're an AI first SEO operator brand identity. It's got some of my brand colors coming through that I've used in chat GPT. How I think so pulling that through what I don't tolerate even vague advice or yes man answers. Now, from here, this is great because we've got this information and you could let it go on and on and on, but we could just say, can you add this to my Clawude memory? And what you'll see is Claude is actually going to go and add this information into your clawed memory. It's going to go manage the internal memories within Claude and actually merge and make sure we're not duplicating anything in there as well. So when you go into your clawed memory, so if we go into settings in claude here and we go down into capabilities and then generate memory from chat history, you'll see all of this memory that we've been building and adding that context which is now being siloed in with chat GPT's one as well. So that's how you're going to extract stuff from the project into your main memories. Now I do also have a prompt here as well that is going to take us a step deeper. This is going to make it populate the memory with even more knowledge and precision. So you can see this is our complete migration prompt here that really breaks down the task into identity projects, knowledge bases, decision-m patterns, so much stuff in here for it to go through. So you could take this prompt and then run it in that project to get more information. But I also want to show you another smart way of actually building this knowledge base out using Claude Co-work. If you've been using Claude Co-work a little bit, then you'll love this method. If not, this is a great introdu in introduction to the sort power that you can use co-work for. Now, in the desktop version of Claude, you have more options. At the top here, you'll see you have chat, you have co-work, and you have code. We're going to stick in co-work for a minute because this allows Claude to actually work with files directly on your computer and actually use your computer for you, which is pretty cool. Now once you get it set up, you can actually add files from your computer that you want Claude to work with. So you can see I've given it access to that data download export that we had from chat GPT. So it can see all that information and use it as a knowledge base. Or the other option is we can actually upload files from this as well. So I'm actually going to give it my chat.html file history that we downloaded. So it has that directly in the chat. And you'll see that it doesn't have that 31 megabyte upload limit. We can upload much bigger files into co-work to start working with. So, we've got all of that chat history loaded up in here. And it's even got access to my full data export as well. Now, from here, we can take this massive complete migration prompt, take that over to Co-work, and now it can start actually running this cognitive migration architect analysis task right within Co-work. So, we'll hit let's go and we'll let it get to work. Now, you might be thinking, why don't we just start using this Elliot? Well, number one, you might not be comfortable using co-work yet if you've never used Claude before. And two, it doesn't interlink so well with clawed memories. Co-works using all the files on your computer and its inbuilt skills. It can't use things from projects. It can't use your chat GPT memory and it can't so easily commit stuff into the memory. But you'll see here the file size 65 megabytes is quite substantial. So it needs to take its time going through strategic section patterns and information. So it's going to go through this, but it's actually going to deliver a one pager report on everything we want to know about me and my business and everything I've done within chat GPT. The other thing you could do from go work is just start working with your chat GBT files as the knowledge base for your chats. Although that leaves you slightly confined to just this area of Claude. Now once co-work has actually finished extracting everything. This is you're going to get we're going to be provided with a complete cognitive migration analysis for us to go through. Now this is really cool because it gives us identity profiles. I'm a volume operator, a systems thinker, timeconscious. It's got all my cognitive thinking styles, communication preferences, as well as going on and on into what I do for my business and all the different projects I've been working on. It even at the bottom has got usage instructions for Claude. This isn't a person who needs handholding. They leverage when they ask for something. Assume high confidence, provide structure, optimize for reuse. This is so cool to actually extract all of this stuff from 23,512 conversations I've had with chat GPT. Now the problem came in trying to add this into my general claude knowledge. When I said can you add this information into my clawed memory, it was building a persistent memory file on my desktop for Claude co-work to use from this information. You can see it's creating a local file on my computer. But I didn't actually want this. I wanted it in my main area of claude chat. So, what you'll have to do is, so the easiest way to deal with So, the easiest way to get this into your main clawed memory is again to copy this out, go back into the main chat, and either place it in a chat or ask it to add it to memory. Or we could even go down onto the sidebar here, open up our settings, go into our capabilities, open up memory from my chats, which you'll probably see is quite similar. It even has more work context for me here, personal context. But we could actually paste, we could actually say, "Hey, I have this memory extraction from chat GPT. Can you merge it into my clawed memories?" And then just paste that information in there and run that and add it into your memory of Claude and it will go and update all of your memory
🎁 FREE Download: Complete ChatGPT to Claude Migration Guide and Prompts inside the AI Recipe Vault: https://eliotprince.com/ai-recipe-vault/ How to switch from ChatGPT to Claude. Should you use ChatGPT or Claude? That's the wrong question. Claude and Anthropic are absolutely killing it right now — Opus 4.6, Co-work, Claude Code. If you're running a business, Claude is hands down the #1 choice. But here's the problem: You've spent months (maybe years) building up ChatGPT. Custom instructions, chat history, workflows that actually work. And the idea of starting from zero? That's what's stopping you. Here's what we're solving today: I'm going to show you how to transplant your entire ChatGPT brain into Claude — your history, your context, your memory — in under 15 minutes. Without losing anything. In this video you'll learn: - How to export your complete ChatGPT history (2 min) - How to upload it to Claude Projects (even if it's massive) - How to populate Claude's memory system with your key context - The "cognitive migration prompt" that extracts patterns you'd never spot - The Claude Co-work method (fastest option for power users) - How to verify everything transferred correctly By the end: You'll have Claude running with full context from ChatGPT — better than before.