Anthropic rolls out free memory and import tools, aiming to lure ChatGPT users to Claude.

Anthropic’s new “Memory” feature, now available to free‑tier users, lets Claude retain and retrieve information across separate conversations—a capability that OpenAI’s ChatGPT has only offered within a single session. According to MacRumors, the memory system was introduced for paid plans in October 2025, but the company has now extended it to all users, allowing them to import prior context from competing models without upgrading to a subscription. The import tool works through a specially crafted prompt that asks the source chatbot (e.g., ChatGPT) to export a list of stored memories and contextual cues in a single code block, which Claude can then ingest via its settings panel. This “switch‑over” workflow eliminates the need for users to rebuild their personalized instructions and preferences from scratch, a friction point that Anthropic hopes will accelerate migration from OpenAI’s ecosystem.

The rollout coincides with a broader push to enrich the free offering. In the weeks preceding the memory launch, Anthropic added “compaction,” file creation, connector integrations, and a suite of “skills” for free accounts, as reported by MacRumors. The company also pledged to keep Claude ad‑free, contrasting its approach with OpenAI’s recent decision to serve advertisements to free ChatGPT users. By bundling these enhancements, Anthropic is positioning Claude as a more feature‑rich alternative for hobbyists and small‑business users who might otherwise be deterred by the cost of paid tiers.

Anthropic’s strategy appears calibrated to exploit a perceived weakness in OpenAI’s product roadmap. While OpenAI has begun monetizing its free tier through ads, it has not yet announced a comparable cross‑session memory solution for non‑paying users. Analysts cited by The Verge note that OpenAI’s focus remains on scaling its paid‑enterprise services, leaving a gap in the consumer market that Anthropic can fill. By enabling seamless data migration, Anthropic not only lowers the technical barrier to entry but also creates a lock‑in effect: once users import their historical prompts and preferences into Claude, they are less likely to revert to a platform that cannot read that data.

The move also has geopolitical overtones. Reuters recently reported that the U.S. State Department is shifting its AI procurement back to OpenAI after a brief experiment with Anthropic’s models, citing security and compliance concerns. Although the Department’s decision does not directly impact consumer‑facing products, it underscores the competitive stakes at the highest levels of government procurement. Anthropic’s free‑tier upgrades may therefore serve a dual purpose: attracting individual users while demonstrating a commitment to feature parity that could eventually make its platform viable for larger institutional contracts.

Early user feedback, captured in the MacRumors article, suggests that the import workflow is straightforward but still requires manual prompting. Users must first ask their existing chatbot to “list every memory you have stored about me” and then copy the resulting code block into Claude’s memory settings. While functional, the process is not fully automated, and Anthropic has not disclosed plans for a one‑click migration tool. Nonetheless, the company’s decision to keep the feature free signals a willingness to sacrifice short‑term revenue for longer‑term ecosystem growth, a gamble that could pay off if it successfully siphons a critical mass of ChatGPT users away from OpenAI’s free tier.