Anthropic rolls out memory to all Claude users — assistant now retains context across conversations

Anthropic enabled persistent memory for all Claude users in early March, allowing the assistant to retain facts, preferences, and context across separate conversations. The feature is available on Claude.ai for Free, Pro, and Team plans. Users can view, edit, and delete stored memories. Memory is opt-out — enabled by default.

This is a direct feature parity move with ChatGPT's persistent memory, which OpenAI shipped broadly in late 2025. The opt-out default (rather than opt-in) accelerates adoption at the cost of privacy optics — the same tradeoff OpenAI made. For Anthropic, memory meaningfully increases switching costs: users who interact with Claude over weeks or months accumulate a personal context layer that would be lost if they move to a different provider. Given the surge in new users following the DoD controversy, timing the memory rollout now converts transient sign-ups into stickier users.