Claude Managed Agents at a glance
Anthropic published 'The components of agent memory' on November 6, 2026, detailing Claude dreaming: an offline pass in which Claude re-reads past session transcripts to correct its memory notes, merge duplicates and build wikis; one dream fixed a wrong note that had trapped Claude in Pokémon Red.
Ask your AI about this story
Ask what changed in Claude Managed Agents and Dreams and what to check before using it in your business. Explain how close this is to something I could actually use, and what is still unproven.
What changed with Claude Managed Agents
Anthropic published 'The components of agent memory' on November 6, 2026, a research post that defines Claude dreaming and the memory system behind Claude Managed Agents and its Dreams feature. Dreaming is an offline pass, run while an agent is idle, in which Claude re-reads its memory notes against the original session transcripts, corrects mistakes, merges duplicates, builds structured representations such as wikis, and decides what future sessions should load first.
Anthropic tested what models choose to remember in Pokémon Red, comparing memory folders at about 10,000 steps into the game: Claude Sonnet 3.7 wrote 413 memory files across 225 folders, 977,930 characters in total, and earned one badge, while Claude Fable 5 kept 10 files totaling 43,406 characters and earned four badges. In the same game, Claude had wrongly recorded that the PokéCenter stood on the east side of Mt. Moon; a dream checked the note against the session transcript, caught the error and rewrote it, and with the corrected note Claude re-entered the cave and reached Cerulean City, which the run with the uncorrected note never did.
Dreaming also builds new representations: working from an internal corpus of about 350,000 files, roughly 420 million tokens, from Stainless, the software company Anthropic acquired in May, Opus 5 agents used 128 dreaming sessions to build and correct a linked wiki, and scores on Anthropic's 25-question evaluation grew from 13% to 48% of graded criteria as the wiki grew. With search and the wiki combined, scores rose from 54% to 61% and used about 20,000 fewer tokens per question.
For developers, the same mechanism ships as Dreams in Claude Managed Agents, documented by Anthropic as a research preview that requires requesting access. A dream takes one existing memory store plus 1 to 100 past sessions and produces a new, reorganized store; the input store is never modified, so teams can review the result before using it. Anthropic's documentation states that Dreams are not eligible for zero data retention or HIPAA coverage.
Why Claude Managed Agents matters for businesses
Why it matters: an agent that keeps its own memory accurate gets more dependable the longer it works for a business, because each new session starts from notes checked against what actually happened instead of stale or wrong ones. A business running Claude agents on support, research or operations can use Dreams as scheduled maintenance for its agent fleet: duplicates and outdated facts get cleaned up between sessions, your team stops re-briefing the same corrections, and fewer mistakes repeat across customers.
Get AI News by email
Verified AI stories, each with its official source, on Monday and Thursday mornings. Add breaking news to hear about the biggest ones as soon as we publish them.
Every story links its official source
Free, in English or Spanish
Unsubscribe with one click, any time
Sources for this story
- The components of agent memory (opens in a new tab)Official
Anthropic · anthropic.com
- Dreams — Claude Platform Docs (Managed Agents) (opens in a new tab)Official
Anthropic · platform.claude.com
Last verified: . IA Compa summarizes public announcements for business readers. We are not affiliated with the companies mentioned. Always confirm details, pricing, and availability with the official source.
- #anthropic
- #claude
- #claude-dreaming
- #agent-memory
- #claude-managed-agents
- #dreams
- #ai-agents
- #memory-stores
