Anthropic explains agent memory components and dreaming for Claude
AIAnthropic describes how Claude keeps memory notes between sessions and uses a process it calls dreaming to review transcripts, correct wrong notes, and build wikis. Its evaluations compare memory-writing rules, the Pokémon Red and CL-Bench tasks, and a wiki built from 350,000 files, where scores on 25 questions rose from 13% to 48% as dreaming scaled from two sessions to 128.
AIWhy it matters.
The post separates memory capture from offline consolidation, with evaluations showing how note quality and wiki structure change agent performance.
Loading source posts…
Source details
Anthropic Research · anthropic.comPublished · added here
Original title: The components of agent memory