Anthropic Adds Free Memory Feature and Import Tool to Lure ChatGPT Users to Claude

[MacRumors]

Anthropic is aiming to lure customers from ChatGPT and Gemini with a new memory import tool that's available to free users as of today. Conversations and memories from other AI providers can be imported into Claude, so new users will not need to start from scratch.

Users can import preferences and context from an AI service like ChatGPT using a prompt that Anthropic prepared for other chatbots. The prompt instructs the AI to prepare a list of memories and context from past conversations, with formatting that's easily digestible by Claude.

I'm moving to another service and need to export my data. List every memory you have stored about me, as well as any context you've learned about me from past conversations. Output everything in a single code block so I can easily copy it.

Claude has an input box for the exported data, and the information is added to Claude's memory. Claude has supported memory for paid plans since October 2025, but there was no option for free users to take advantage of the feature. With a memory feature now available for switching, users will not need to sign up for a paid plan to move their AI instructions and preferences to Claude. Anthropic says that memory will remain an option on the free plan.

Enabling memory and then importing information into Claude can be done through the Memory section of Claude's settings.

Last month, Anthropic took advantage of ChatGPT's decision to start displaying ads for free users, promising to keep Claude ad-free. Anthropic then added new features for free users, including options for compaction, creating files, using connectors, and accessing skills. Along with memory, those options were previously limited to paid subscribers.

As of now, Claude is the top free iOS app in the App Store, a spot normally held by ChatGPT. Anthropic has been in the news over the last week for its inability to reach an agreement with the U.S. government over AI use, and the subsequent supply chain risk designation it received.

Note: Due to the political or social nature of the discussion regarding this topic, the discussion thread is located in our Political News forum. All forum members and site visitors are welcome to read and follow the thread, but posting is limited to forum members with at least 100 posts.

Article Link: Anthropic Adds Free Memory Feature and Import Tool to Lure ChatGPT Users to Claude

 

[TheDailyApple]

Huh, I wonder if competitors will take steps to prevent easy exports like that.

Personally the feature holds no value for me as if I’m using an AI, I want as fresh an instance as possible so that its output is less targeted at pleasing me specifically. I usually disable memory features for that reason sticking to a single generic overall instruction on behavior.

 

[jz0309]

well, they lost a government contract so now the money needs to come from consumers ... Beware ...

 

[tonmischa]

I have never used Anthropic's Claude before. But I like that they stand by their core values (even if it means losing business).
Take note, Tim. That's what having a spine looks like.

 

[Gengar]

I only use Gemini since I have so many Google services anyway, but good on Anthropic for making Claude the superior option for the masses.

 

[KurtWilde]

well, they lost a government contract so now the money needs to come from consumers ... Beware ...

The exact reason OpenAI jumped to take their place.

 

[Keymaster]

Huh, I wonder if competitors will take steps to prevent easy exports like that.

Personally the feature holds no value for me as if I’m using an AI, I want as fresh an instance as possible so that its output is less targeted at pleasing me specifically. I usually disable memory features for that reason sticking to a single generic overall instruction on behavior.

This should give you the capability to do both. So, you can save conversations you have had and when you want to work on that again load it up so that you start where you were, but when you want to take on something new you can start fresh. I can see where that would be useful, I have seen AI bring old stuff into a conversation when changing a topic to something new.

Claude is the only AI I would use for coding, it's managed to solve problems that other AIs couldn't when I ran some tests, and it's actually shown me a new (and maybe better) way to do a couple things. I literally had to tell ChatGPT the answer to issues before it could finally give me something that was relevant, and it would try to correct software based on code that was commented out. It's not exactly confidence building that the military is going to use it, or really any AI...AI screws up so much that it can cause a lot of problems if it's not used judiciously, and if it's being used to evaluate intelligence or select targets a lot can go very wrong.

 

[Blackstick]

Claude is the only LLM I use to code. Bash and powershell scripts in particular, it writes like butter. While not full on development, there's a lot of syntax and boilerplate code nobody wants to burn time on.

Anyway. Claude's work is almost always right first time. Used to be I'd spend 3-4 hours writing and debugging a script. That timeframe has shrunk to 10 minutes and the machine's output is tighter, more resilient, has better self healing/error recovery, less race conditions and just overall better optimized. It's really quite extraordinary.

 

[It’s always something]

Anthropic’s principled stand was enough to win me over. I use Claude now when I want to use an AI.

 

[sevoneone]

Gemini doesn't seem willing, lol:

 

[ProbablyDylan]

Am I the only one that dislikes "memory" features? I don't need a chatbot remembering stuff from other conversations.

 

[turbineseaplane]

Am I the only one that dislikes "memory" features? I don't need a chatbot remembering stuff from other conversations.

The companies themselves sure love it I’ll bet.

 

[sevoneone]

The exact reason OpenAI jumped to take their place.

Altman had just secured another $110 Billion in Venture capital before all of this happened. Investors that have never seen a single dollar in return drop $110 Billion a couple days before Altman becomes Hegseth's best friend. Makes you wonder about the entire situation and story... Now they can get return on investment with taxpayer money by overcharging the government just like every defense contractor ever.

 

[CarAnalogy]

I’ve been an AI enthusiast as far as the chatbots but at this point it seems they are actually dragging each other down.

Each one keeps imitating the other and getting worse because of it. And they are all pushing way more notifications at me in their apps.

But the worst part is they are all getting dumber. Too many guardrails and poisoned sources and anti bias biasing.

 

[CarAnalogy]

Am I the only one that dislikes "memory" features? I don't need a chatbot remembering stuff from other conversations.

As creepy as it is, it’s the only way it can ever become truly useful. But I agree the ship has sailed (if it were ever in port) on being able to trust these hosted chatbots with memory.

The ******tification cycle started early and it’s already being used against us.

Mods seriously add that word to the allow list. It’s in the dictionary at this point.

 

[CarAnalogy]

I have never used Anthropic's Claude before. But I like that they stand by their core values (even if it means losing business).
Take note, Tim. That's what having a spine looks like.

Enjoy it while it lasts. I hope I’m wrong but I don’t think that’s a spine, I think that’s opportunism.

It’s what the very last shred of functional capitalism looks like, where they still feel like they need to look like they care what users want.

 

[wikiverse]

Am I the only one that dislikes "memory" features? I don't need a chatbot remembering stuff from other conversations.

ChatGPT is completely useless and can't even remember something that was discussed 3 messages ago. It also cannot read the existing thread that you're in.

If they don't have long term memory, it's like working with someone with Dementia - you just end up repeating yourself and explaining things again and again.

 

[TFrank2]

ChatGPT is completely useless and can't even remember something that was discussed 3 messages ago. It also cannot read the existing thread that you're in.

If they don't have long term memory, it's like working with someone with Dementia - you just end up repeating yourself and explaining things again and again.

The memory settings/experience has improved a bit, but nowhere near perfect. It can definitely be frustrating at times

 

[rp2011]

I have never used Anthropic's Claude before. But I like that they stand by their core values (even if it means losing business).
Take note, Tim. That's what having a spine looks like.

Hell yeah, not even a fan of Dario as every time he opened his mouth in interviews was always cringeworthy. But now I am certainly a fan. Way to go. We are private citizens. FREE. He doesn't have to answer to lord Vader.

 

[klasma]

well, they lost a government contract so now the money needs to come from consumers ... Beware ...

Anthropic raised $30 billion in the last round, and the government contract was only $200 million.

 

[kiv.atso]

Unless they reduce the price or increase the amount of tokens, this is way too expensive in comparison with ChatGPT. I use Claude at work and it’s great, but for personal I don’t see myself paying £100/month to be able to get the level of usage I get from ChatGPT for £20. No way.

 

[locksmack]

I just made the jump from ChatGPT to Claude. The memory transfer was easy, though I was hoping it would give me actual transcripts of previous chats that could then be seen as chats in Claude, but no (that was probably too ambitious).

I do wonder if the gained users will outweigh the gov contract for their bottom line (and what it means for OpenAI).

 

[jz0309]

Anthropic raised $30 billion in the last round, and the government contract was only $200 million.

"only $200M" now, in the long run they will not get any new gov contracts as well as private sectors who have any ties to gov. And some enterprises will follow ...
I'm not saying what they did was wrong, but the biz implication is huge

 

[klasma]

Hell yeah, not even a fan of Dario as every time he opened his mouth in interviews was always been cringeworthy for me. But now I am certainly a fan of Dario. Way to go. We a private citizens. FREE. He doesn't have to answer to lord Vader.

He’s a bad speaker/interviewee and is too optimistic about AI improvements, but he and his co-founders are principled people who actually care about ethics.

 

[klasma]

as well as private sectors who have any ties to gov.

That’s not correct, at least legally. If the “supply chain risk” thing goes through (it hasn’t been formally confirmed yet, and they have good chances to overturn it in court), it only blocks other companies from using Anthropic specifically in the context of government contracts; all other uses by the same companies are unaffected.