On Thursday, Anthropic rolled out its Memory feature to all paid subscribers as part of its ongoing effort to enhance Claude’s capabilities for enterprise users.
The Memory feature enables the generative AI assistant Claude to retain ideas and contextual information accumulated across sessions. For researchers, this means the system can progressively accumulate source code over time, while developers benefit from preserved coding environment configurations.
According to Anthropic, users can view what Claude remembers and edit those memories using natural language commands.
Anthropic has also introduced additional Claude capabilities for Pro and Max plan subscribers, including web search for real-time information, voice mode for hands-free collaboration, and research tools for investigative tasks.
Memory isn’t entirely new for Anthropic—it was previously launched last month for Team and Enterprise plan users. Moreover, it’s not a groundbreaking innovation, as similar memory features already exist in OpenAI’s ChatGPT and Google’s Gemini, noted Mark Beccue, an analyst at Omdia, part of Informa TechTarget.
“To some extent, the differences might be quite subtle,” Beccue said, adding that broadly speaking, it’s difficult to distinguish between these offerings.
In a press release, Anthropic highlighted transparency as a key differentiator. Claude users can review the AI’s stored memories, instruct it to focus on specific topics, or tell it to forget certain data points. However, OpenAI also allows users to manage ChatGPT’s memory by editing stored information or disabling references to saved memories.
In an email to AI Business, Anthropic emphasized that its Memory feature was designed with enterprise needs in mind. Claude organizes memories by project, keeping contexts separate. The model focuses on learning user workflows rather than merely collecting isolated facts, the company said.
Benefits and Challenges
William McKeon-White, an analyst at Forrester Research, described the new capability as a valuable addition.
“Claude has proven more effective than other models at retrieving specific details and assets,” McKeon-White said. “This gives users another tool to deepen contextual understanding and solidify their reliance on the platform.”
He also pointed out several challenges associated with the new feature, particularly how it functions in conflicting or ambiguous contexts.
“How will it scale in enterprise deployments? Memory as a feature is foundational for building ‘smarter’ AI—but it’s a stepping stone in that journey, not the final destination,” he noted.
The Memory feature complements Anthropic’s existing offerings, including chat history and incognito chat modes.
Max-tier subscribers gained access to Memory today, while Pro users will receive access in the coming days.
This marks Anthropic’s second feature-related announcement this week. On Monday, the company launched Claude Code on the web, enabling users to delegate coding tasks directly from their browser. The web-based Claude Code is currently in a research preview phase.
