Anthropic's Claude finally gets memory, catching up to ChatGPT and Gemini

Anthropic announced the memory feature for its Claude AI last month for Team and Enterprise users. While ChatGPT rolled out memory back in early 2024, Anthropic is finally opening it up for individual paid users as well, starting with the Max subscribers and Pro users over the next two weeks.

Memory allows Claude to remember user interactions and project-specific contexts over sessions, improving continuity and productivity.

A standout feature in Claude's memory design is its project-specific approach, where each project maintains its own "memory space", keeping relevant conversations and details focused and separate. This is different than ChatGPT's implementation of memories, where it gets relevant data from all the chats instead. Of course, with Claude, users can ask it to prioritize certain memories or to “forget” specific past information.

Claude also has an option for incognito chats, where it won't save the conversation to memory or your chat history, and nothing will stick around after you close the conversation.

Claude has been getting a few other skills as well recently, quite literally, with the rollout of "Claude Skills," where users can create custom instructions in markdown files, tailored for specific tasks such as data analysis or document generation. Anthropic also announced the availability of Claude Code on the web and on iOS, where users can push code, submit PRs, and a lot more directly from their browsers and smartphones.

To get started with memories, users can ask Claude questions like “What were we working on last week?” to see what Claude remembers across their existing chats and connected tools. Users can also import memories from other AI tools, such as ChatGPT, and even export memories from Claude, by following these instructions.