Perplexity Launches Brain, a Self-Improving Memory System That Builds a Context Graph of an Agent’s Work and Learns Overnight

Most AI memory remembers the user. It stores your preferences, your tastes, and your role. Perplexity is taking a different path. Today, Perplexity launched Brain , a self-improving memory system for its agent product, Computer . Brain does not focus on remembering you. It remembers what the agent did. That reframes what memory in AI is for.

Brain is a self-improving memory system. It builds a context graph of the work Computer performs. At set intervals, such as overnight, Brain reviews that graph. It then teaches itself how to do the work better. The idea is straightforward. The more work you do, the more efficient Brain makes your Computer. Brain is rolling out today to Perplexity Max and Enterprise Max subscribers in Research Preview.

Perplexity frames memory along two axes . The first is what the memory is about. The second is what the memory is for.

Traditionally, AI memory has been about the user. It stores preferences, tastes, working styles, contacts, and role. Its purpose is engagement. It helps you feel more engaged with the agent. Brain takes the other path. Its memory is about the agent’s work. It remembers what worked, what failed, and what corrections got made. Its purpose is performance. Perplexity calls helping the agent get better the most important purpose of memory.

Brain forms a living context graph for Computer. The graph is traceable. It helps Computer understand the user’s world and learn from their work. The context layer takes the form of an LLM wiki. That wiki is automatically loaded onto the agent sandbox. Its pages reflect the ideas, people, projects, and other elements in a user’s world. Computer can traverse this web of personal information.

The Brain system updates the wiki incrementally overnight. It synthesizes the user’s sessions, connector results, changes in source documents, and corrections made. That refreshing context gives Computer a stronger signal on what to do and where to look.

Brain also shows its work. Every memory entry links back to the session, file, or source it came from. That traceability matters for debugging and trust.

Brain gets better as you use Computer. Agents learn the projects, connectors, artifacts, and other sources that lead to the best outputs. They also learn from their mistakes. They remember when a user has made a correction. They remember when a source was a dead end. That results in fewer turns, fewer model calls, and better outputs. This feedback loop is what makes Brain continuously self-improving. Perplexity team frames current token usage as an investment in more efficient token usage later.

Perplexity shared early measurement results from its own testing.